Jonathan Kanter, Assistant Attorney General for the Department of Justice Antitrust Division, recently gave a speech condemning the use of the consumer welfare standard in antitrust cases. While current application of the standard is often imperfect, his proposed replacement is conclusory and unhelpful.
The head of the Department of Justice’s Antitrust Division, Jonathan Kanter, made a recent speech adopting the Neo-Brandiesian position that in antitrust cases the consumer welfare standard should be jettisoned and replaced with a “competition and the competitive process” test. I quite agree with him and the Neo-Brandiesians that the way that courts have applied the consumer welfare standard has led to uncertainty and underenforcement. But I think his approach raises serious concerns of legal theory and strategy.
Concerns of Legal Theory
Kanter is right that antitrust law protects “competition and the competitive process”.[1] But the consumer welfare standard was never an alternative to that legal test. It was only a method to resolve deep ambiguities about what “competition and the competitive process” means, and Kanter offers no alternative to resolve them.
To take a simple concrete case, suppose a merger between two firms in a ten-firm market makes the merged firm a more efficient and vigorous competitor. Does that decrease “competition” because it eliminates one competitor or increase it because we now have more vigorous competition among the nine that remain? Nothing in the “competition and the competitive process” test helps provide an answer.
Kanter defines “competition” to mean “rivalry”, i.e., the existence of competitors, and “the competitive process” to mean the “freedom to choose” among those competitors. Given those definitions, his test could mean one of three things, none of which is helpful.
First, Kanter’s definitions are consistent with the position that antitrust law bans any conduct that reduces the number of competitors and market choices. But then antitrust law would ban two plumbers, in a market with 1000 plumbers, from forming a partnership to offer better services. He can’t mean that. That sort of approach would limit our economy to atomistic competition between sole proprietors in a way that would massively reduce our productivity and impede our economic liberty to collaborate with others in efficient ways.
Second, Kanter’s definitions are consistent with the position that his test protects only the existence of some competitors and market choice. But then the antitrust laws would allow any conduct other than a merger or cartel that results in a 100% monopoly. After all, even a merger that creates a market duopoly still results in rivalry between two competitors and the ability to choose between then. He can’t mean that either, given that his avowed goal is to increase antitrust enforcement.
The third alternative is that he means that antitrust law bans conduct that does not leave “enough” competitors and choice. But that makes the test wholly conclusory. It provides no help in deciding what is enough and when the freedom to choose among better competitors is preferable to the freedom to choose among a greater number of them. Kanter provides no metric for determining such issues. One could resolve such ambiguities by asking whether the number and nature of competitors and choices in that market sufficed to protect some substantive antitrust goal, but Kanter specifically rejects that approach. Indeed, he rejects using any supplemental standard to resolve ambiguities in his test, insisting the analysis “starts and ends” with “competition and the competitive process” test. As such, his test amounts to a conclusory I-know-it-when-I-see-it test. This hardly provides the “clear, administrable” test that Kanter promises.
Kanter offers various defenses. First, he argues that all legal tests require the exercise of judgment in close cases. Fair enough. We all know what “baldness” means even if we cannot define precisely how many hairs one needs to lose before one turns bald. But Kanter’s test is not just vague at the margins; it is conclusory in almost all cases. In the 10-to-9 merger case, it cannot even determine whether subtracting one firm increases or decreases competition. It is like a baldness test that cannot even tell us whether subtracting a hair makes a person balder or not.
Second, Kanter argues that the consumer welfare standard is also uncertain because “if you ask five antitrust experts what the consumer welfare standard means, you will often get six different answers.” This sort of “whataboutism” argument provides no response to the concern that his own test is utterly conclusory. Further, although he is right that experts disagree about the meaning of the consumer welfare standard, none of those differing meanings is conclusory; instead, each provides affirmative guidance on how to resolve the otherwise ambiguous cases. Moreover, the differences are exaggerated in that the different definitions of the consumer welfare standard generally lead to the same legal conclusions in those ambiguous cases, including focusing on the same factors to resolve the 10-to-9 merger case. To the extent there remain differences, the solution to that problem would be for the agencies and courts to adopt the version of the consumer welfare standard that is best. Or to adopt some different standard that can help resolve the ambiguities in a better way, if one feels such a standard exists. Either is preferable to jettisoning all substantive standards in favor of a conclusory I-know-it-when-I-see-it test.
Third, Kanter argues that the consumer welfare standard is substantively bad and leads to underenforcement. Again, this amounts to a form of whataboutism. If the consumer welfare standard is bad, the solution is to improve it or replace it with a better standard for resolving ambiguities, not to demand using no substantive standards at all by insisting that the analysis “starts and ends” with the conclusory competition and the competitive process test.
Moreover, Kanter’s substantive objections to the consumer welfare standard are either misplaced or capable of being handled in ways other than jettisoning that standard. One of his substantive objections is that “some versions [of the consumer welfare standard] . . . assert the antitrust laws were never intended to protect our democracy from corporate power, or to promote choice and opportunity for individuals and small businesses.” I am unaware of any versions that take that position, but if so, the answer is to use one of the versions that does not.
“None of this is to say the consumer welfare standard is necessarily the optimal standard. Perhaps it is not, though I have not so far seen a better one.”
The position I have instead seen taken by advocates for the consumer welfare standard is that the problem with applying a “corporate power” or “choice and opportunity” test is that they, too, are conclusory in that they provide no basis in ambiguous cases for determining how much corporate power is too much and how much choice and opportunity is too little. Thus, one needs some standard for resolving those ambiguities, which the consumer welfare standard can provide.
There is something to that position, but it does have the unwarranted effect of giving no weight to political concerns about corporate power or to freedom concerns about choice and opportunity, even though they clearly were important purposes for enacting the antitrust laws. Even if courts lack any method for weighing such political or freedom concerns in an administrable way in individual cases, the law instead could take them into account categorically by providing that, whenever the economic effects of conduct are ambiguous, conduct should be presumed unlawful if it concentrates markets or restrains market choices. This shift in the burden of proof would have systematic effects that reduce corporate power and increase market freedom, but that still leave firms able to engage in conduct that efficiently benefits consumer welfare if they could prove it really does so. It would also greatly reduce the antitrust underenforcement that Kanter rightly decries.
Another of Kanter’s substantive objections is that the consumer welfare standard leads to uncertainty and underenforcement because it reduces antitrust cases “to econometric quantification of the price or output effects of the specific conduct at issue,” which is not administrable by courts or predictable by businesses. As a result, he claims, “Too often, it leads us to focus on estimating data to the third decimal point for statistical models detached from the competitive realities actually playing out in the markets.”
I would challenge Kanter to cite any antitrust case that has ever turned on whether plaintiffs or defendants are right about the third decimal point in a statistical model. I have never heard of such a case. Still, he has a point that requiring case-by-base econometric analysis of the effects of conduct creates uncertainty and underenforcement. But this is not a consequence of the consumer welfare standard itself. It is rather a consequence of insisting that it be applied in a way that requires open-ended, case-by-case, all-things-considered analysis of whether consumer welfare is enhanced or harmed by any specific conduct.
This current way of applying the consumer welfare standard not only creates uncertainty, but also undermines deterrence. It imposes enormous litigation costs on plaintiffs. It also frequently requires econometric proof of matters that cannot be established given data limitations. Those data limitations are sometimes inherent, but are often worsened when courts do not order the needed production of data from defendants, other market participants, and often other markets. All too often, defendants successfully fight the production of data, and then win on the grounds that the lack of data prevents the needed econometric proof, much like Abraham Lincoln’s lament about the “man who murdered his parents, and then pleaded for mercy on the grounds that he was an orphan.”
However, there is nothing in the consumer welfare standard that requires either using such an open-ended case-by-case approach or demanding econometric proof even when it cannot be supplied. Courts could instead use the consumer welfare standard to adopt the rules or presumptions best calculated to advance consumer welfare given actual data limitations. For example, a recent merger retrospective by John Kwoka found that the consumer welfare effects of mergers were predicted better by using certain structural presumptions than by using open-ended case-by case consideration of whether the merger enhanced consumer welfare. If one agreed with his results, that would provide a strong basis for using a structural presumption rather than open-ended case-by-case assessment of the welfare effects of each merger.
Indeed, the irony for the Neo-Brandeisians who want to jettison welfare standards is that, during the actual Brandeisian era, antitrust enforcement was weak and ineffectual. The reason was that it depended on conclusory assertions that sometimes being “too big” was bad, which resulted in isolated and mercurial enforcement. This “meant little deterrence of anticompetitive conduct both because enforcement was unlikely and also because it was unclear just what firms were supposed to do to avoid enforcement.” Under a conclusory test in which liability is effectively random, the threat of such liability has little effect on behavior, other than imposing a random tax that discourages being in business at all.
This Brandeisian era of weak and ineffectual enforcement was ended by the appointment of Thurman Arnold to head the Antitrust Division in 1938. “Arnold made clear that (unlike his predecessors) he had no problem with businesses being big as long as their conduct was efficient and lowered consumer prices. This gave firms a far clearer and more desirable signal about how to modify their behavior.” Further, based on the economics, he pressed to replace prior conclusory tests about whether conduct restrained competition with per se rules that prohibited certain forms of conduct like horizontal price-fixing, even when the alleged cartelists had no plausible market power. The effect was to greatly increase antitrust enforcement and the use of rules. The consumer welfare standard is thus perfectly consistent with the adoption of more rules and presumptions.
Kanter’s final substantive objection to the consumer welfare standard is that it “has a blind spot to workers, farmers, and the many other intended benefits and beneficiaries of a competitive economy.” I think this objection is mistaken. To be sure, some misinterpret the consumer welfare standard to permit anticompetitive harm to upstream suppliers of labor or other inputs. But as I have pointed out, those interpretations are mistaken because the consumer welfare standard just says that antirust will not protect competitors from conduct that affirmatively improves consumer welfare. See Elhauge, U.S. Antitrust Law & Economics 208-209 (3d ed. 2018). Any anticompetitive harm to upstream suppliers will suppress upstream output, and if that has any effect on downstream output, it will be to reduce it and thus harm consumer welfare. Id.
Concerns of Legal Strategy
Leaving aside what the best test is in theory, is Kanter’s approach a sound legal strategy? I think not. Take the desire of Kanter and other Neo-Brandeisian enforcers to increase antitrust enforcement to protect labor and farmers from anticompetitive monopsony power. Because of their commitment to the proposition that labor and farmers cannot be protected under the consumer welfare standard, they are effectively admitting that, in any case in which the competition and competitive process test raises ambiguities, labor and farmers cannot win the case under the consumer welfare standard that has been adopted in six Supreme Court cases and countless lower court decisions.[2]
The only way they can win such a case, then, will be to persuade a court to overrule the consumer welfare standard. But no trial or appellate court will overrule a standard that has been enshrined in six Supreme Court cases, so (given the length of antitrust cases) they will not be able to win such a case using this strategy during the Biden administration, even if he serves two terms. Nor is it likely that the Supreme Court will overrule its own adoption of the consumer welfare standard, given ordinary standards of stare decisis and the current composition of the Court. In contrast, a litigation strategy that adopts my position that the consumer welfare standard does protect against anticompetitive harm to upstream suppliers, such as labor and farmers, seems far more likely to succeed because it does not require any court to overrule Supreme Court precedent.
Kanter’s argument to the contrary is that in one of those Supreme Court cases, Reiter v. Sonotone, 442 U.S. 330, 343 (1979), the reference to the Sherman Act being a “consumer welfare prescription” was dicta because the holding was simply that consumers were one sort of party with antitrust standing, and the opinion acknowledged that others injured by anticompetitive conduct also had antitrust standing. Fair enough, though it is awfully persuasive dicta for any lower court because the full statement was that “Congress designed the Sherman Act as a ‘consumer welfare prescription’”, id., meaning that it was a reading of statutory purpose that should help resolve statutory ambiguities. Moreover, for reasons I just noted, such a consumer welfare purpose is perfectly consistent with giving standing to non-consumers injured by the anticompetitive conduct as well.
In any event, there are at least four other Supreme Court cases for which the adoption of a consumer welfare standard cannot be dismissed as dicta because it was necessary to their holdings. In Brooke and Weyerhaeuser, it was necessary for the holdings that, respectively, below-cost pricing and overbidding without subsequent recoupment should be allowed as procompetitive because they benefit consumer welfare while they exist without harming them in the long run, rather than condemned as anticompetitive because they inefficiently impede rival competition while they exist. In Leegin, it was necessary for the holding that the prior per se rule against vertical minimum price-fixing agreements should be overruled because such agreements could benefit “consumer welfare” with higher service levels, rather than condemned per se because they restrain free market choice and could eliminate some dealers. Finally, in Jefferson Parish, it was necessary for the holding that the quasi per se rule against ties with tying market power but no substantial tied foreclosure share should be retained because such ties could increase the exploitation of that market power in a way that harmed consumer welfare, rather than allowed on the ground that such ties did not harm “competition” because they did not increase the degree of tied or tying market power. This last holding shows that a consumer welfare standard can produce more enforcement than intuitions about whether competition has been decreased.
More generally, to the extent that Kanter’s speech indicates that the Department of Justice will no longer try to show harms to consumer welfare, but rather will rely on conclusory assertions about whether the conduct harmed “competition and the competitive process”, I think that is unlikely to succeed in the courts. The defendants in such cases will offer evidence that consumer welfare was not harmed, and if the government offers no contrary evidence, the issue will be treated as undisputed, and the courts will rule for defendants under the existing consumer welfare standard. Moreover, even if the courts could be persuaded to change the consumer welfare standard, they will likely reject an approach, like Kanter’s, that “starts and ends” with a conclusory assertion of whether the conduct harmed “competition and the competitive process” on the grounds that such an approach adopts an arbitrary and capricious standard of liability.
None of this is to say the consumer welfare standard is necessarily the optimal standard. Perhaps it is not, though I have not so far seen a better one. But the competition and competitive process test provides no guidance in many cases, and in such cases jettisoning the consumer welfare standard in favor of standardless conclusory judgments is unsound both as a matter of legal theory and legal strategy.
[1] A caveat that applies throughout this piece: whether conduct harms “competition and the competitive process” is generally a necessary but not sufficient element under the antitrust laws. For example, Sherman Act §1 also requires an agreement, Sherman Act §2 monopolization claims also require monopoly power and that the conduct have some causal connection to that power level, and Clayton Act §7 also requires an acquisition of stock or assets.
[2] Reiter v. Sonotone Corp., 442 U.S. 330, 343 (1979) (“Congress designed the Sherman Act as a ‘consumer welfare prescription.”); NCAA v. Bd. of Regents of Univ. of Okla., 468 U.S. 85, 107 (1984) (same); Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 15 (1984) (“the consumer” was the one “whose interests the [Sherman Act] was especially intended to serve.”); Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 127 S. Ct. 2705, 2713 (2007)(equating an “anticompetitive effect” with being “harmful to the consumer” and “stimulating competition” with being “in the consumer’s best interest.”); Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 224 (1993) (allowing below-cost pricing without recoupment because it benefits “consumer welfare”); Weyerhaeuser Co. v. Ross-Simmons Hardwood Lumber Co., 129 S. Ct. 1069, 1077 (2007) (allowing overbidding that results in below-cost downstream pricing without recoupment because it benefits “consumer welfare”); John B. Kirkwood & Robert H. Lande, The Fundamental Goal of Antitrust: Protecting Consumers, Not Increasing Efficiency, 84 Notre Dame L. Rev. 191, 219-24 (2008) (collecting cases). They would still be able to win cases where the buyer conduct is per se illegal because such cases raise no ambiguities that require resort to these sorts of tests for resolution, but those cases could already be brought and were.
Correction: Actually, eight Supreme Court cases have adopted the consumer welfare test, with the other two being the most recent substantive antitrust opinions by the Supreme Court. See Ohio v. American Express Co., 138 S. Ct. 2274, 2284 (2018) (equating an “anticompetitive effect” with being “harmful to the consumer” and “stimulating competition” with being “in the consumer’s best interest”); NCAA v. Alston, 141 S. Ct. 2141, 2151 (2021) (same). Thanks to John Kirkwood for pointing that out. This makes it even less plausible that a litigation strategy that depends on rejecting the consumer welfare standard can succeed in the courts.
Read more about our disclosure policy here.
