The Roberts Court and the Limits of Antitrust
I’ve just finished a draft of a paper for an upcoming conference on the Roberts Court’s business law decisions. Volokh blogger Jonathan Adler, who directs the Center for Business Law and Regulation at Case Western, is organizing the conference. The other presenters are Adam Pritchard from Michigan (covering the Court’s securities decisions), Brian Fitzpatrick from Vanderbilt (covering pleading standards), and Matt Bodie from St. Louis University (covering labor and employment). My paper discusses the Roberts Court’s antitrust decisions.
I am not the first to analyze the Roberts Court’s antitrust jurisprudence. Both Josh and Einer Elhauge have written terrific papers on the themes underlying the Court’s antitrust decisions. Josh has argued that the Court’s antitrust jurisprudence reflects Chicago School thinking; Elhauge contends that it’s more aligned with the Harvard School. While I’d probably side with Josh in that debate (mainly because I think the “new” Harvard School, having correctly jettisoned the old Structure-Conduct-Performance paradigm, moved so starkly in Chicago’s direction that it should really be called Chicago-Lite), I need not take sides. That’s because the unifying theme I identify in the Roberts Court’s antitrust decisions, one acknowledged by both Josh and Elhauge, is common to both the Chicago and Harvard schools. That theme is a recognition by the Court that antitrust is an inherently limited body of law that must be constrained in its reach — even to the point of allowing some undesirable conduct — if it is to do more good than harm.
Lots of folks are upset by the constraints the Roberts Court has imposed on antitrust. As Josh noted on this blog, Erwin Chemerinsky has complained that the Court’s antitrust decisions exemplify a “sharp turn to the right” and systematically “favor business over consumers.” Chemerinsky, of course, is no antitrust expert. But even well-respected members of the antitrust community have sounded a similar refrain. For example, William Kolasky, a former Deputy Assistant Attorney General in the Antitrust Division of the U.S. Department of Justice and an associate editor of the ABA’s Antitrust Magazine, recently (though before the Court’s most recent antitrust decision) wrote that “Our Supreme Court, especially under the leadership of Chief Justice John Roberts, seems equally intent on cutting back on private enforcement. … This record led Antitrust to ask in its last issue whether the Supreme Court’s recent antitrust decisions represent ‘The End of Antitrust as We Know It?’”
In my paper, “The Roberts Court and the Limits of Antitrust,” I argue that anti-consumer/pro-business/”radical shift” meme the Chemerinskies and Kolaskies of the world assert is wrong and reflects a fundamental misunderstanding of the inherent limits of the antitrust enterprise. Below the fold, I describe the paper.
PART ONE: I begin with an explanation of antitrust’s limits. (That discussion, based largely on Judge Easterbrook’s insights, will be quite familiar to regular TOTM readers. It’s a point Josh and Geoff have emphasized a number of times.) Antitrust tries to make life better for consumers by ensuring vigorous competition. It does so through a standard-based body of law in which (1) courts must decide, after the fact, whether a business practice is “unreasonable” (either an “unreasonable” restraint of trade or an “unreasonably exclusionary” practice); (2) private plaintiffs have a liberal right to sue, ensuring judicial review of a great many challenges; and (3) damages are automatically tripled, causing firms to avoid even kosher practices that are close to the line of illegality (or are simply novel) and that may therefore be deemed unreasonable by nonexpert judges or, worse yet, juries.
This system is costly. First, simply reaching a decision about legality frequently requires significant economic investigation and is therefore costly. Second, mistakes are inevitable, meaning that some bad acts will go unpunished (so market power, with its attendant allocative inefficiencies, will result), and some output-enhancing practices will be falsely condemned (so that productive efficiencies are squandered). Taken together, these decision costs and error costs constitute the limits of antitrust. Those limits are inexorable. One cannot reduce decision costs without increasing error costs, and attempts to avoid error costs related to false negatives will generate more false positives and vice-versa. Given these inevitable and inexorable limits, courts ought to craft antitrust liability rules in a way that minimizes the sum of decision and error costs. I term such an effort a “decision theoretic” approach; Josh and Geoff, emphasizing that error costs are generally the biggest factor in the calculation, call it “error cost analysis.”
PART TWO: Having set forth the limits of antitrust and prescribed a decision-theoretic approach for dealing with those limits, I next march through the Roberts Court’s antitrust decisions (leaving out the one Robinson-Patman decision, given that that statute is expressly not focused on consumer welfare) and explain how each coheres with the decision-theoretic perspective. So, for example:
Leegin (holding that minimum resale price maintenance (RPM) is not per se illegal but must be evaluated under the rule of reason): RPM can be pro- or anticompetitive. Economic theory suggests that the prerequisites to its procompetitive uses will be satisfied far more frequently than those for any anticompetitive harm. Moreover, empirical evidence indicates that most instances of RPM are, in fact, procompetitive. Error costs will therefore be reduced by jettisoning automatic condemnation in favor of an approach that requires the plaintiff to establish actual harm (or, at a minimum, the prerequisites to such harm). While decision costs may rise somewhat with a rule of reason approach, the many loopholes under the per se rule (Colgate, Monsanto, Business Electronics) — exceptions that were created because the per se rule was economically unsound — actually made that rule pretty difficult to implement. So, while decision costs will rise somewhat, the rise won’t be as great as one might initially suspect and, in any event, will be dwarfed by the reduction in error costs.
Weyerhaeuser (holding that predatory bidding claims are subject to the same “Brooke Group” requirements as predatory pricing claims): An input buyer may “overbid” for many procompetitive or competitively benign reasons. Moreover, as Keith Hylton has observed, the prices created by unfettered bidding generate socially useful information. Accordingly, attempts to regulate input overbidding that does not result in below-cost pricing of one’s output are likely to create high error costs. Moreover, the decision costs of any alternative rule (such as the Ninth Circuit’s direction that liability may result from paying “a higher price than necessary” for one’s inputs) would be tremendous.
I’m not going to go through the rest of the decisions here (lest this already-too-long blog post become ridiculous and no one read my paper!), but you get the idea. Suffice it to say, each of the decisions — even American Needle, as Josh and Judd Stone have observed — coheres nicely with an approach aimed at minimizing the sum of decision and error costs.
PART THREE: I end by making predictions about what the Roberts Court, if it sticks to the decision-theoretic path it has trod thus far, will do on some antitrust issues likely to arise in the future. With respect to tying, I predict that the Court will abandon the quasi-per se rule in favor of a rule of reason that is focused on the degree of tied market foreclosure. Such an approach will not sit well with post-Chicago scholars, like Einer Elhauge, who contend that non-foreclosure-causing tie-ins may injure consumers by facilitating price discrimination and the extraction of consumer surplus. But, while tying-induced price discrimination may occasionally cause static efficiency losses, it more frequently enhances static efficiency, and permitting it generally increases dynamic efficiency by encouraging sellers to develop new and differentiated products on which they may earn supracompetitive profits via price discrimination. Attempts to identify and condemn welfare-reducing instances of tying-induced price discrimination would involve both high error costs and high decision costs. The Roberts Court is therefore likely to limit the tying prohibition to tie-ins that may expand the defendant’s market power by foreclosing significant marketing opportunities for tied market rivals. (Indeed, as I have previously noted, the Independent Ink Court acknowledged that tying-induced price discrimination is perfectly consistent with competitive markets.)
My other predictions relate to loyalty rebates and bundled discounts. With respect to the former, I predict that the Court will hold that such price cuts are legal unless they run afoul of Brooke Group (i.e., the discounted per-unit price is less than the seller’s cost, and there is a likelihood that the seller could recoup its losses via supracompetitive pricing). Because the price cut inherent in a loyalty rebate represents a competitive bird in the hand, and because any equally efficient rival that was willing to lower its price to the level of its cost could avoid being excluded by a loyalty rebate, the Court will likely reject post-Chicago scholars’ arguments for complex liability rules that could capture extremely rare instances of anticompetitive harm stemming from above-cost loyalty rebates but would, along the way, discourage vigorous price competition. With respect to bundled discounts, I predict that the Court will, at a minimum, approve of a safe harbor for bundled discounts that are above-cost under the discount attribution test (i.e., the price of the competitive product is still above cost after the entire discount is attributed to that product). It will thus side with the Ninth Circuit in the circuit split created by the Ninth Circuit’s PeaceHealth decision and the Third Circuit’s LePage’s decision.
I’m currently tightening the paper, filling in cites, etc., so I haven’t yet posted the piece to SSRN. If anyone would like to review a very early draft, just email me — I’d be happy to share, and I’d much appreciate any comments.