Solving the Robinson-Patman Act / Consumer Welfare Conundrum
The FTC is bringing the Robinson-Patman Act back. That's a bad thing for competition. How bad? And what can be done about it? Part 1 in a 3 part series explaining the Act and what comes next.
Welcome back to Competition on the Merits! Last issue focused on what a second President Trump antitrust program might look like and how it might differ from the first time around. Thanks to all of you for subscribing.
This week we get into the weeds of antitrust law and economics and kick off a 3-part series on the FTC’s promised return of the Robinson-Patman Act, what it means and how a modern wave of FTC RPA cases might play out in court.
“Robinson-Patman is ineffective when evaluated both in terms of its narrow, protectionist objectives, and in terms of its benefits to the welfare of society as a whole. The greater the business community’s compliance with Robinson-Patman, whether as a result of voluntary action or vigorous public or private enforcement, the greater its deleterious impact upon competition.”
DOJ Report on the Robinson-Patman Act (1977).
The conflict between the Robinson-Patman Act (RPA) and competition has been well understood for a very long time. The DOJ understood it well by its exhaustive tome evaluating it in 1977. The bipartisan Antitrust Modernization Commission did as well when it recommended its wholesale repeal in 2007. Nearly all antitrust scholars, both lawyers and economists, have concluded that RPA enforcement harms competition and does nothing to protect small business. But its here — courtesy of the FTC — and apparently not going anywhere but to Article III courts. So let’s talk talk today about the RPA itself and explore the conundrum of RPA enforcement — that is, why does more enforcement harm competition and consumers? Is that conclusion correct? Is that interpretation of the Act required under the language of the statue? I’ll conclude today’s post with an argument that the Supreme Court (or maybe even an appellate court) is likely to put an end to the RPA / consumer welfare conundrum. But let’s start with some groundwork.
The RPA walks and talks like a normal antitrust law. But it is not one. Some bemoan this fact. Others celebrate it. Most of that divide is explained by whether one views the tension between the RPA’s protection of competitors rather than competition as a bug or a feature of the Act. But it is an important starting point for understanding the Act. And it is important to understand the nature and structure of the RPA to answer a set of interesting questions about its promised, and looming, enforcement by the FTC. At least one of those questions is whether the FTC will win or lose the upcoming cases. But there are others.
The RPA’s structure and language does not offer much help in providing simple answers to these questions, as we shall see. It is also the victim of poor drafting. I am not the first to offer that criticism. Academic commentators, legislators, and courts have all recognized the same. My own favorite flavor of this critique comes from the D.C. Circuit in Boise Cascade (1988): “The need for clarity, obviously a desideratum for any body of law, has been evidenced by judicial recognition that Robinson-Patman is not to be viewed as an act of Congressional schizophrenia, an anti-competitive island situated in an otherwise turbulent sea of pro-competitive efficiency and maximization of consumer welfare, the hallmarks of the Nation’s antitrust laws.”
But let’s start with something simpler. The goal of this issue is to unpack and simplify a few different aspects of the RPA - a challenge in its own right because the statute is horribly drafted. The poor drafting has resulted in some impressive intellectual gymnastics in lower court decisions that has no doubt made it even more difficult to predict how future courts will apply the RPA. It is also remarkably important to separate out RPA history, including precedent, before and after the Supreme Court’s 1993 Brooke Group decision, which affirmed that the RPA’s primary-line injury requirement (more on that soon) was to be interpreted in harmony with the Sherman Act’s monopolization provisions. Brooke Group was not the Court’s first nod at resolving the tension between earlier cases that ground RPA liability in evidence of harm to a competitor - the kind of harm that would arise because of competition between firms and to the benefit of consumers. But it marks an important shift in RPA history with the Court and is important to keep in mind.
If we are going to figure out the answers to important questions like:
Is RPA enforcement good or bad competition policy? Should the RPA be repealed, as recommended by the bipartisan Antitrust Modernization Commission (among others)?
Will the FTC win its upcoming cases?
What sort of arguments and evidence will the FTC and defendants make in those cases?
It is important to begin with some basics and, in particular, a deeper understanding of the tension between the economic foundations underlying the rest of the antitrust laws and the RPA’s penchant for liability based upon harm to competitors regardless of the impact on competition and consumers. Is the tension grounded in legislative history? The statute? I think a useful way to unpack some of the issues is starting with a very different question: “Where does consumer welfare fit in the RPA?”
THE BASICS
There are some great resources to read on the history of the Robinson-Patman Act. Anybody interested in a more complete treatment should get their hands on Areeda-Hovenkamp, or this oldie but goodie from Richard Posner published by AEI, or the various ABA monographs on the subject.
This is meant only to be a very simple primer aimed at providing just enough to highlight a pretty interesting conundrum created by the RPA and the Supreme Court cases interpreting it when it comes to economic evidence.
The RPA contains six basic provisions: Sections 2(a) through 2(f). We will spend most of our time with 2(a) - which prohibits a seller from discriminating in price between different buyers. 2(b) lays out affirmative defenses for price differences that are cost-justified or offered to “meet” a competitor’s lower price. 2(c) covers commissions and brokerage fees. 2(d) and 2(e) prohibit a seller from providing or paying for promotion or advertising allowances on a differential basis to competing buyers. 2(f) introduces buyer liability under the RPA by making it unlawful “knowingly to induce or receive” a discriminatory price otherwise prohibited by the Act. In other words, if Walmart knowingly asks for or receives Coca-Cola for a better price than other retailers receive and it knows that better price is not cost-justified or offered to meet a competitor’s price, it may run afoul of 2(f).
We will mostly focus on 2(a) and 2(f). For now, you can think of 2(a) as the statutory basis for cases against product suppliers (think, Coca-Cola) offering differential prices to retailers. 2(f) is the other side of the coin, the suit against the supermarket retailer “knowingly inducing or receiving” the discriminatory price from the supplier.
At what level of the market harm is alleged to occur is also an important feature of RPA cases. For example, if the harm is horizontal, that is - the harm occurs at the same level as the alleged RPA violator, it is referred to as primary-line injury. For example, Coca-Cola offers differential prices to retailers – with some getting favored (lower) prices and others higher prices. Pepsi sues Coca-Cola alleging the discounted prices offered to Coca-Cola’s favored customers violate RPA 2(a) and cause harm to Pepsi – that is, a horizontal rival (Pepsi) of the supplier offering the discounts (Coca-Cola) is harmed. That is a primary-line 2(a) case.
Primary line injury is the economic injury most antitrust types are familiar with. In order for Coca-Cola’s discounts to weaken Pepsi as a rival and cause harm to interbrand competition - that is, competition at the soda supplier level – it would have to be the case that the discounts either caused Pepsi to exit or raised Pepsi’s costs such that Coca-Cola could raise market prices profitably. We have standard tests in antitrust for theories of harm to interbrand competition. In Brooke Group, the Supreme Court said primary line theories of RPA harm must satisfy a test similar to the one set forth for Sherman Act Section 2 predatory pricing cases. There are differences--whether a plaintiff must show a “dangerous probability”of successfully harming competition versus only a “reasonable probability,” and so forth. But the critical point is that the Supreme Court has already held that the required currency of the plaintiff’s burden in a primary line RPA case is consumer welfare, not merely harm to rivals.
As you might imagine, plaintiffs do not often win primary line cases. Or rather, they win them about as often as plaintiffs win predatory pricing claims under Section 2 of the Sherman Act.
But there is another theory of harm: secondary line injury. OK, there is tertiary too. But we are not going to talk about it for now. Secondary line injury occurs at the level of the allegedly unlawful discounter’s customers. For example, if Coca-Cola offers larger discounts to Walmart than it does to other retailers competing with Walmart, and the FTC sues Coca-Cola alleging the discounts harm competition between Walmart and the “disfavored” retailers, that is secondary-line injury. To be clear, some courts use shorthand here: suppliers are primary line, and retailers are secondary line. But I think the economic logic is a little more robust here (for reasons we will discuss): horizontal harm is primary line (i.e., Coca-Cola is the defendant and allegedly harms competition in the soda supplier market) and vertically-related harm is secondary (i.e., Coca-Coca allegedly harms downstream retailers).
Secondary-line injury is where things get interesting, and not only because that is the direction the FTC’s cases are heading. Secondary-line cases are where this awkward tension between harm to rivals and harm to competition really stands out and has persisted in the cases, even after Brooke Group resolved that tension for primary-line cases. Consider the theory of harm in a 2(f) claim, let’s say against Walmart (whom Commissioner Bedoya has singled out in some speeches). The theory of harm would be that Walmart’s allegedly unlawful conduct (knowingly inducing or receiving discounts) harms a horizontal rival in the retail market. That sure sounds primarily-line to me. But it is an open question in these cases - when they inevitably come - whether courts will follow the economic reasoning of the claims or take comfort in the courts that have reflexively assigned “all things retail” to secondary-line.
Think of the 2(a) cases against suppliers as a good warm up exercise for an agency that has little to no experience litigating RPA cases. But the 2(f) cases will eventually come. Once one has marshaled the evidence necessary to establish a 2(a) case against a supplier, it is a simple modification to bring a 2(f) case alleging that a powerful buyer “knowingly induced or received” the differential discounts that were at issue in the 2(a) case and that those discounts resulted in competitive harm. Would that harm be primary or secondary-line? Again, many of the cases reflexively treat that harm as secondary-line because it occurs at the retail level. But that conclusion is at odds with the economic logic of treating horizontal competitive harm as primary-line injury.
BACK TO THE CONUNDRUM
So what is so special about the RPA? Why do critics get so worked up about it? Here is why: secondary-line cases generally have only required a plaintiff to show that the disfavored buyer (retailer) is harmed in order to satisfy its competitive injury requirements. Let’s say that again: harm to a rival in the form of lost sales to the favored purchaser have been sufficient to establish injury. In other words, the only burden a plaintiff faces in establishing competitive harm in the secondary-line RPA context is showing that when Coca-Cola gives a big discount to Walmart and a little discount to the mom and pop store – that the mom and pop store loses sales to Walmart.
Some (potentially important) technical requirements, details, and caveats live here. For example, the favored and disfavored retailers have to compete against each other in the same relevant market in order for the requisite injury to exist. This requirement is another place where modern economic analysis in the form of diversion analysis and market definition tools can find their way into RPA cases. But again, ALL a plaintiff has to show here is that the retailer with the bigger discount and lower prices enjoys some diversion of sales from the higher price rival. Additionally, the plaintiff must prove that the favored retailer knew (or should have known) that the discounts it received were unlawful under 2(a)--that is, they were not cost-justified or given to meet competition.
When students are taught in law school that “antitrust protects competition, not competitors” – someone ought to raise their hand and say “Except the RPA!” because the RPA does exactly that. A successful plaintiff need only show that the disfavored rival was harmed in the form of lost sales to the favored retailer. A small aside here. Private plaintiffs also have to show standing by proving “antitrust injury,” that is, the plaintiff’s injury must be one that the antitrust laws were designed to prevent. That doctrine does deter some private lawsuits, though it has been fairly loosely applied in secondary-line RPA cases. Nonetheless, for our purposes, the FTC does not face that requirement. The bottom line is that a government plaintiff can always, and a private plaintiff can quite often, establish RPA liability without a single iota of evidence that the defendant’s discounts harmed competition and consumers. A demonstration that sales shift from the higher price retailer to the lower price retailer are sufficient. The RPA protects competitors at the expense of competition. Mind blowing. We are back to the conundrum.
Indeed, sometimes the plaintiff does not EVEN have to prove the lost sales. The so-called Morton Salt inference (a case decided long before Brooke Group) permits – but does not require – courts to infer the requisite harm in situations where a substantial price difference between competing buyers persists. Again, this inference is permissive. It is not required. But it allows courts to find the requisite harm with even less evidence - just the observation of a substantial and persistent price difference.
I hope we have established by now how different secondary-line RPA cases are from the rest of the antitrust universe. But why on earth is the RPA so different? RPA is not a separate antitrust law. It is an amendment to the Clayton Act. As Professor Hovenkamp has lamented, the secondary-line RPA provisions are “irritating to almost anyone who is serious about antitrust.” Indeed. But why has the RPA been so effective at shutting out the consumer welfare standard – or any economic proof that the competitive process was harmed by the allegedly unlawful discounts?
One sensible hypothesis is that the statutory language is different. That argument coupled with appeals to the legislative history which reveals various clear statements that the RPA is targeted at large retailers are usually relied upon to comfort folks that are serious about either antitrust or statutory interpretation that the unique features and harmful consequences of the RPA for consumers are the way things are meant to be. Perhaps this is all to be expected in a statute drafted by the Wholesale Grocers’ Association. But perhaps not.
If one reads discussions and scholarly commentary and Commissioner speeches about the RPA one might be persuaded that it is and will forever be an outlier. Consumer welfare-centric antitrusters might be left only with hope of repeal because there is no chance for an RPA interpretation in secondary-line cases that is consistent with the rest of our antitrust laws (and thus, the consumer welfare standard). So we are told.
I want to challenge that notion. And I want to challenge it with what I acknowledge is a provocative argument. One that is an uphill road that might well lose in district court given the tremendous amount of ground covered in lower court cases (though, mostly before Brooke Group) and the easy route for a lower court to follow the path set before it by other lower courts that have largely accepted the notions that: (1) secondary-line harm to competition only requires evidence of harm to a rival; and (2) therefore, benefits to consumers or arguments that the allegedly unlawful discounts increased competition rather than “injuring, destroying, or preventing it” are irrelevant to analyzing an RPA claim. But I’m not so sure the argument below would not prevail with a supra-majority of the Supreme Court right now.
Let us start with the statute. Given the astronomical differences between the rest of the antitrust universe and the RPA secondary-line cases, one is probably expecting major differences here. Prepare to be underwhelmed.
The statutory test is whether:
the effect of such discrimination may be substantially to lessen competition or tend to create a monopoly in any line of commerce, or to injure, destroy, or prevent competition with any person who either [i] grants or [ii] knowingly receives the benefit of such discrimination, or [iii] with customers of either of them…
That is really three different tests. Two of three should be familiar. “To substantially lessen competition” and “to tend to create a monopoly in any line of commerce” are precisely the language in the Clayton Act we apply to mergers – where there is no doubt that proof that the conduct creates market power to the detriment of consumers is required. Whether or not the language of the statute specifies it – the first two tests appear to read on primary-line injury. Sometimes the requisite effect is inferred from market share statistics. But the plaintiff clearly faces the burden of proof to demonstrate the transaction alleged to violate Section 7 of the Clayton Act creates market power that would not exist but for the transaction and that as a result, consumers will be made worse off. If one is looking to statutory language for a decent explanation of why standard economic analysis does not fit into the RPA cases – the first two prongs sure do not help.
What about the third test? It is a bit different, right?
“To injure, destroy or prevent competition with any person who either grants or knowingly receives the benefit of such discrimination…”
What about this language? Notice first what it does not say. Surely Congress understood when it amended the Clayton Act the difference between harm to a competitor and harm to competition. Nonetheless, the RPA speaks of discounts that “injure, destroy, or prevent competition with any person who either grants or knowingly receives….” It would be a simple thing to prohibit using discounts to injure or destroy a competitor. The RPA does not do that. The RPA language itself expresses interest in competitive conditions – in the case we have been discussing the inquiry would be whether Coca-Cola’s favored discounts to Walmart injure, destroy, or prevent competition between Walmart and disfavored retailers. The RPA speaks in terms of assessing whether competition at the retailer level has been injured, destroyed, or prevented.
If you ask Industrial Organization economists about various methods to assess the state of competition in a particular market – and in particular, whether competitive intensity has changed in a particular market in response to some conduct – you might well get a variety of answers but nearly all would mention analyzing what happens to prices and output in that market. Most economists would start with that answer. A very conventional method for assessing claims that particular conduct has harmed competition in the rest of the antitrust universe is to ask what price and output would be but for the allegedly anticompetitive conduct. If the conduct causes prices to fall and output to rise, the conduct has increased competition–not harmed it.
Where RPA cases have gone off the rails, as discussed, is that the lower courts have generally answered this question simply by observing (or presuming, a la the Morton Salt inference) that sales flow from the higher price retailer to the lower price retailer. In fact, there are numerous cases interpreting the secondary-line competitive injury requirements to be satisfied by proof that the favored retailer’s lower prices not only resulted in diverted sales from the higher price retailer but also attracted new customers such that output went up! That is – proof that consumer welfare went up, along with output, is a SUFFICIENT CONDITION for RPA liability. There is probably no more clear illustration of the tension between the general antitrust laws – and their fidelity to the consumer welfare standard – and the application of the RPA in practice.
The RPA’s own language does not command this result or the harm to consumers that flows from it.
I started this article by noting what I described as a bit of a conundrum: the RPA is an outlier among the antitrust laws because consumer welfare is totally ignored, at best, in a large subset of cases (e.g. secondary line cases), yet much of the RPA’s own language is identical to Clayton Act language that has been endlessly interpreted to require standard economic analysis of effects upon competition and consumer welfare. The third RPA test also includes words (“to injure, destroy or prevent competition with any person who either grants or knowingly receives the benefit of such discrimination…”) that seem to directly suggest that the court look at the impact on competition rather than individual competitors. Despite the statutory language, secondary-line RPA cases have remained a “consumer welfare-free zone.”
Why have the secondary-line RPA cases been so successful in evading the consumer welfare standard and goals of the general antitrust laws? Why have they been so successful in evading economic evidence and proof of economic harm to consumers? And is it permanent? How long will secondary-line RPA cases remain a safe space for hand-waving about impacts upon competition rather than actually analyzing it?
I offer an optimistic view. Antitrust analysis changes over time. Substantive legal rules and procedural rules evolve. This is part of the beauty of the “common law” antitrust approach. Conduct assessed as per se unlawful for a century switches to rule of reason. Courts change liability rules and respond to changes in economic evidence and legal thinking. The RPA is NOT immune from those trends. Decisions after Brooke Group look different than decisions before it.
Progress has already been made. The RPA is already evolving. Slowly. But it is. The Supreme Court clearly already agrees that the RPA language does not preclude an interpretation that requires harm to the competitive process and the creation of market power. Indeed, the Supreme Court has interpreted the RPA to require exactly that sort of proof in primary-line cases but refused so far to extend that logic to secondary-line cases.
Will the evolution reach secondary-line cases like those the FTC is poised to bring? The good news is that the Supreme Court has certainly flirted with the idea of harmonizing secondary-line RPA cases with the rest of the antitrust laws. In Reeder-Simco, the Court did its best to try to harmonize secondary-line RPA cases with the general antitrust laws:
Interbrand competition, our opinions affirm, is the "primary concern of antitrust law. The
Robinson-Patman Act signals no large departure from that main concern. Even if the Act's text could be construed in the manner urged by Reeder and embraced by the Court of Appeals, we would resist interpretation geared more to the protection of existing competitors than to the stimulation of competition . In the case before us, there is no evidence that any favored purchaser possesses market power, the allegedly favored purchasers are dealers with little resemblance to large independent department stores or chain operations, and the supplier's selective price discounting fosters competition among suppliers of different brands. By declining to extend Robinson-Patman's governance to such cases, we continue to construe the Act consistently with broader policies of the antitrust laws.
Is there any place for economics in secondary-line cases at all? And I do not mean the sort of economic analysis that would merely measure and trace diversion from the disfavored buyer to the favored buyer. These are important measurements – but I mean, is there a place for standard economic analysis focused upon the impact of a business practice on competition and consumers in secondary-line RPA cases?
I think the answer is unequivocally yes. And if I were defending these cases I would assert as much directly and invoke the RPA’s own language to do it. I would do so understanding I may not prevail with that argument at the district court (though, maybe!) but believing that the chances in appellate courts and the Supreme Court are considerably greater.
Let’s think about this concretely with a hypothetical that will probably look like each of the FTC’s RPA cases to come. You are representing the supplier in a 2(a) case OR the large, favored retailer in a 2(f) case. Ignore for now the cost justification and meeting competition defenses – which you will obviously also raise. Ignore also the economic analysis you offer to support your defense that the favored and disfavored retailers are not in the same relevant market. Recall, secondary-line theories of harm require the favored and disfavored purchaser to compete in the same market. By the way – recent FTC practice in grocery retail cases will provide plenty of ammunition for arguments that many retail outlets fall outside the scope of the relevant market where Walmart is the favored retailer. For example, the FTC has truncated grocery markets to exclude premium supermarkets (e.g., Whole Foods) as well as discount grocers (e.g., Aldi) in litigated cases. These defenses and arguments narrow the scope of relevant purchases and impact damages. But let’s get back to the point.
If you are defending these cases and you believe the economic reality is such that the large discounts to the favored retailer ultimately result in lower prices to consumers and higher output, is that a cognizable defense? If you are representing Walmart – you almost certainly believe that your client’s low prices, which flow from greater discounts arising from intense competition among suppliers, result in substantial benefits to competition and consumers. Does the RPA really exclude that evidence in the secondary-line context? How do you get that evidence in before the federal judge assigned to your case? What does it look like? What is the legal hook? Is consumer welfare in the traditional sense relevant under the RPA?
I think you will hear a lot of talk that it is not. I do not think that is correct. Not by a long shot. I’m writing in longer form on this issue as we speak to take a deeper dive into some of the cases, but we can spell out the basic elements of the argument here.
First, a plain reading of RPA 2(a) contemplates an analysis of harm to consumers. How so? The statute requires–even in secondary-line cases–a plaintiff to show that the differential prices “injure, destroy, or prevent competition with any person who either grants or knowingly receives the benefit of such discrimination…” An obvious method to show that competition between the favored retailer (e.g. Walmart) and the disfavored retailers (e.g., other grocery retailers) has increased rather than been “injured, destroyed, or prevented” is to show that market prices went down and output went up.
Second, the RPA cases focusing on a shift in sales from the disfavored retailer(s) to the favored retailer simply do not preclude courts from engaging in a deeper and more accurate analysis of competition (the actual word in the statute) rather than merely looking at harm to competitors.
Third, focusing on price and output as a method of assessing whether competition has been substantially lessened, or injured, destroyed, or prevented, as opposed to enhanced or promoted is natural, consistent with other antitrust cases using similar language, and consistent with sensible economic principles underlying the modern antitrust laws.
Fourth, as I believe Professor Hovenkamp has pointed out, the RPA is part of the Clayton Act. It is not a separate antitrust law. The Clayton Act contains a glossary of terms. One of those terms is competition. Another is monopoly.
Fifth, not only has the Supreme Court already taken exactly this step in primary-line cases (e.g., Brooke Group) by aligning its test with the predatory pricing cases, but it has also strained itself to harmonize the RPA with the other antitrust laws. In Reeder-Simco, the Court acknowledges interbrand competition as the primary concern of the antitrust laws and does its darndest to keep the RPA within the family of antitrust laws that are at least somewhat consistent with that goal. Here, a reading of secondary-line cases consistent with interbrand competition as the primary goal of the antitrust laws focuses on competition at the retail level–that is, competition among retailers for consumers–and the outcome of that competitive process. The Supreme Court has already expressed demand to “construe the Act consistently with broader policies of the antitrust laws,” and these cases will put an unavoidable choice in front of the Court: whether to harmonize RPA secondary-line cases with the rest of the modern antitrust laws focused upon consumer welfare or to admit that the RPA will continue on as an outlier aimed at subsidizing particular competitors.
I suspect the Court will continue on the path it has already started. When it does, it will have the RPA’s own language, its own decisions interpreting the RPA, sound economics, and good policy on its side. Parties defending these suits ought not accept the limited sandbox of secondary-line cases–particularly in district courts–that defendants sometimes appear constrained to offer. The argument sketched out here provides a roadmap that I believe maximizes the opportunity for RPA defendants who are offering consumers lower prices and enhancing competition to win. We might even get some good case law out of it.