The Robinson-Patman Act Is Even Worse Than You Think: Affirmative Defenses and Unintended Consequences

This is the final installment of our 3 part RPA series. Read about two key defenses and why a RPA comeback might lead to more vertical integration. Just like the Neo-Brandeisians drew it up, right?

If you read our earlier Competition the Merits entries on the FTC’s promised reinvigoration of the much-maligned Robinson-Patman Act, perhaps I have persuaded you that it is a really bad idea. I hope so. If not, at a minimum I hope I’ve at least convinced you to be open to the idea that the longstanding and bipartisan objections to the RPA are based upon sound economics and that the energetic return of the RPA is something to watch closely.  

Today’s objective is to convince you it is even worse than you think it is.  Let’s get to it.  

To review, in our previous two entries we focused on two different issues.  First, we explored the text of the RPA, explaining why secondary-line cases have to this point eluded the consumer welfare revolution, and offered an argument that traditional economic evidence of consumer benefits from discounts have a place in RPA analysis even under existing law.  Next, we debunked a popular myth propagated by enthusiastic RPA supporters and apologists – the myth that there is no economic evidence that RPA enforcement banning discounts to large retailers like Walmart and Costco harms consumers.  Goalpost-moving aside – one would think a responsible standard for public law enforcement should be a standard deviation or two above “there is no evidence this harms people” – we presented evidence from industrial organization economics that the very pricing practices the FTC prosecutes and seeks to ban result in enormous benefits for consumers, disproportionately enjoyed by low-income Americans.  

Today’s third and final entry in our RPA series – final at least until the FTC files its cases, which will merit further discussion on the specific merits – will turn to a new issue.  I want to introduce readers to two RPA defenses: meeting competition and cost justification.  I want to discuss the basics of these defenses because they work a bit differently than traditional antitrust defenses and readers following the FTC cases will want to be familiar with how they work and the type of evidence we can expect to be presented when those cases are filed.  “Conventional” antitrust defenses in all other antitrust contexts – restraint of trade, monopolization, and mergers – generally take the form of defendants presenting evidence that the practice or transaction at issue either does not harm competition as alleged by the plaintiff, or that it actually benefits competition and consumers.  In our earlier entry we discussed how those arguments often played a lesser role in secondary-line injury RPA cases, but also why we suspected the Supreme Court or even appellate courts would be open to “consumer welfare” arguments and why we predict the Supreme Court eventually (and probably because of these cases) will harmonize the RPA with the rest of the antitrust laws.

But these two RPA defenses – meeting competition and cost justification – are more mechanical, technical, and sometimes economically backwards.  The plan for today’s entry is to present the basics of these defenses, and explain a bit about how they are generally applied.  Mostly, I want to use them to raise an issue about the economics of RPA enforcement that is often not discussed.  To be sure, economic criticisms of the RPA are prevalent and well-grounded.  Typically, they take the uncontroversial view that, if you punish conduct (large discounts) that generally results in consumer benefits, you’ll get less of the conduct and consumers will get higher prices.  That critique is well-grounded in economic analysis and empirical evidence as well as common sense.  But RPA enforcement is even worse than that.  Let’s start with the defenses and draw out some of the more pernicious incentives the RPA provides.

The RPA has two primary defenses: (1) the Cost Justification and (2) the Meeting Competition.  This short column is not the place for a treatise on either, but we will spell out the basic elements and issues with the design and structure of these defenses, how they work in practice, and the perverse incentives they create for market participants.   

The Cost Justification Defense

Section 2(a) of the RPA provides that: 

[n]othing herein contained shall prevent differentials which make only due allowance for differences in the cost of manufacture, sale, or delivery resulting from the differing methods or quantities in which such commodities are to such purchasers sold or delivered.

The basic idea of this defense is relatively straightforward.  A seller who can show that its differential pricing is explained only by differences in certain types of costs can immunize itself against RPA liability.  

Somee anticompetitive aspects of the cost justification defense are obvious.  The first is that the RPA allows a defense for a limited set of cost differences.  In the RPA’s own text, only “cost of manufacture, sale, or delivery resulting from the differing methods or quantities in which such commodities are to such purchasers sold or delivered” are cognizable when considering the cost justification defense.

The Cost Justification Defense Imposes Barriers to Entry for Price Reductions

Back to our Coca-Cola example.  Coke sells to thousands upon thousands of different customers.  From Walmart to supermarkets to chain stores and restaurants, the categories of customer are quite diverse.  And even within categories there is substantial heterogeneity.  To justify a price difference among customers, Coca-Cola bears the burden of putting together an expensive and time-consuming “cost study” to demonstrate that the costs of serving the customer was lower, and therefore justified.  This would be expensive enough to do at some aggregated level –  for example, by comparing the costs of serving supermarkets to the cost of serving restaurants.  But, encouraged by the FTC, courts have interpreted the cost justification defense quite narrowly – sometimes requiring the seller to slice the data thinly enough to identify cost differences at the individual customer level.  

Here’s the Supreme Court in Automatic Canteen discussing the Commission’s bizarre position with respect to the level of detail it might require to allow a seller to lower its price:  

... [T]he Commission has not been content with accounting estimates; a study seems to be required, involving perhaps stop-watch studies of time spent by some personnel such as salesmen and truck drivers, numerical counts or invoices or bills and in some instances of the number of items or entries on such records, or other such quantitative measurement of the operation of a business. 

Stop-watch studies needed to lower a price?  The cost justification defense is anticompetitive on its face.  But it is worse in practice.  It serves as a peculiar sort of barrier to entry for lower prices.  An occupational licensing regime for price cutters, if you will.  The seller’s right to compete may be exercised only with the permission of the court.  Permission the courts have generally withheld, with the FTC’s blessing, by refusing to find discounts justified by differences in costs.  

The Cost Justification Defense Deters Discounts

What does this look like in practice?  Let’s say Coca-Cola wants to start selling to a large retailer that has previously been doing business exclusively with Pepsi.  Coca-Cola offers a large quantity discount to the retail chain to get the business and Coca-Cola’s overall output increases substantially.  Because of the large increase in output, Coca-Cola’s average cost of producing soda falls as output expands.  Coca-Cola goes to court and argues it is justified in offering the new customer a lower price because it is able to serve them at lower cost and, in fact, reduces its average costs substantially.  Coca-Cola loses and cannot avail itself of the cost justification defense.  This makes zero economic sense.  But courts have held that because the cost reduction applies to all of Coca-Cola’s output it would be required to share that reduced cost on a pro-rata basis with all customers equally in order to avail itself of the defense.  The unsurprising economic result is that the seller is deterred from offering the discount in the first place.  The RPA defense structure thus raises the cost to the seller of reducing its prices.  It must figure out how to allocate the reduction in cost across all output and all customers.  And then it must offer a discount to each customer that tracks the pro-rata share of that cost reduction. 

Discount to none or discount to all, the RPA says.  The RPA thus discourages discounting by raising the cost of offering a discount.  

The Meeting Competition Defense

RPA Section 2(b) provides the Meeting Competition defense: 

[N]othing herein contained shall prevent a seller rebutting the prima-facie case thus made by showing that his lower price or the furnishing of services or facilities to any purchaser or purchasers was made in good faith to meet an equally low price of a competitor, or the services or facilities furnished by a competitor.

The basic idea is that the RPA provides an affirmative defense for the seller who commits an RPA offense by lowering the price to a particular customer – whether new or existing – to meet the lower price offered by a rival.  Because the RPA by its normal application generates very bad consequences for consumers, it is a reasonable conclusion that exceptions and defenses to the RPA are generally a good thing.  And that is true enough.

But each of these defenses come with some pernicious side effects.  Understanding them requires a quick drive through their requirements.

First, the RPA is a tool of managed, manicured, and planned competition, not vigorous “fight to the death” competition. 

To avail oneself of the meeting competition defense, as the statute says, requires a “good faith” effort “to meet an equally low price of a competitor.”  If Coca-Cola wants to lower its price to Walmart because it believes Walmart is about to give Pepsi its business, the RPA provides Coca-Cola a defense so long as Coca-Cola can show it has a good faith belief that it is meeting a lower price offered by Pepsi.  Now you might be thinking — as a society don’t we want Coca-Cola to lower its price whether or not it knows what Pepsi is doing?  Aren’t all price reductions by Coca-Cola to obtain business – even those encouraged by the uncertainty of NOT knowing Pepsi’s price – good for consumers?  Both answers are unequivocally yes.  Therein lies the problem: the RPA softens price competition by threatening serious penalties for price reductions that do not meet the RPA’s exacting requirements.  

To make things worse, the Meeting Competition defense, as the name implies, only applies to attempts by the seller to “meet” a rival’s low price.  You read that correctly.  Pepsi competes by offering a lower price to a Coca-Cola customer.  Coca-Cola learns of the lower price and decides to, well, compete by offering an even lower price to keep the business.  This is the very essence of competition.  Here, Coca-Cola responding to aggressive competition from Pepsi is actually the sort of price competition that benefits consumers.  But the RPA makes it unlawful unless Coca-Cola meets Pepsi’s competitive strategy by lowering its price to each and every one of its customers.  Competition is good.  Vigorous competition is better.  The RPA discourages it and, perversely, keeps prices higher than they would be without it.  

Second, the RPA Meeting Competition Defense encourages collusion.  

This RPA defense requires the seller to make some investment into knowing and understanding its rival’s prices.  Again, a seller lowering its price must provide evidence that it knows, or at least has a good faith basis for its belief, that a rival has a lower price.  Now, some of you readers are antitrust lawyers.  But you do not need to be an antitrust lawyer to spot the red flag here.  Antitrust is primarily concerned with agreements between competitors on price.  So much so that those agreements are felonies punished with prison time and heavy fines.  And here is the RPA encouraging sellers to potentially communicate with their rivals about price by requiring evidence of its good-faith belief.  

One need not even fully subscribe to Adam Smith’s observation that “[p]eople of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices” to raise an eyebrow at this requirement.  

A line of old cases has expanded the ways a seller can satisfy its good faith requirement from solely actual inter-seller verification of prices, to a broader kind of and quality of information to substantiate a reasonable belief it is meeting a lower price.  Nonetheless, the RPA puts the seller in the position of guessing whether a court will find the basis for its price reduction adequate and provides incentives to insure against that risk by getting better information.  

The Supreme Court has acknowledged this risk a seller who simply wants to lower its price to maintain or to attract customers while avoiding RPA liability faces in U.S. v. Gypsum (1978):

[t]he defense may be rendered unavailable since unanswered questions about the reliability of a buyer's representations may well be inconsistent with a good-faith belief that a competing offer had in fact been made. As an abstract proposition, resort to interseller verification as a means of checking the buyer's reliability seems a possible solution to the seller's plight, but careful examination reveals serious problems with the practice.  … Especially in oligopolistic industries such as the gypsum board industry, the exchange of price information among competitors carries with it the added potential for the development of concerted price-fixing arrangements which lie at the core of the Sherman Act's prohibitions.

On the one hand, a seller who relies on the word of a customer that a rival has lowered its price might be out of luck if the court does not find the customer sufficiently reliable.  (Out of luck to the tune of an RPA violation and trebled damages.)  On the other hand, the discounting seller runs the risk of exposure under Section 1 of the Sherman Act for communicating with the rival seller and getting better information.  

Maybe the seller just keeps the high price and avoids the problem  – not to mention the prospect of antitrust liability and treble damages –  by charging a uniformly higher price.  As discussed in Part 2 of our RPA series, the economic evidence demonstrates that sellers competing for the business of larger retailers with lower prices generates enormous benefits to consumers.  

Third, The RPA is Even Worse Than You Think!  Unintended Consequences of the Reinvigorating RPA Enforcement 

By now I do hope that I’ve convinced you that the RPA is likely to chill discounting and lead to higher prices for consumers – especially low-income consumers.  But I want to end today’s newsletter with a bit of industrial organization economics aimed at highlighting one important unintended (and very likely) consequence of greater RPA enforcement: greater vertical integration.  

The result is more than a little ironic given the Neo-Brandeisian’s are loathe to encourage vertical integration. Indeed, they have celebrated their own hostility toward vertical integration (a practice that largely increases welfare) not only in bringing a handful of vertical merger challenges (not without success) but also adopting new Merger Guidelines that would bring back a 1960s-style “everything is illegal” approach to the transactions.  

Let’s explore the relationship between RPA enforcement and vertical integration by thinking about what contractual relationships generate RPA exposure.  We discussed this in great detail in Part I of our series, the RPA / Consumer Welfare Conundrum, for those just joining us.  The paradigmatic RPA prosecution involves the following characteristics: 

• A seller or distributor who sells its product or products through multiple retail channels (e.g. mass merchandise, supermarkets, convenience stores, etc.);

• At least one of the retailers (buyers) is a large format retailer (Walmart, Costco, Target) that receives larger discounts as a result of its size and business model; 

• A layer of distributors between the seller and retailer responsible for delivery and distribution; 

The first two criteria are obvious enough.  They generate the conditions most likely to create sales at different prices to different retailers that still compete against one another in the same market.  These are important to meeting the prima facie elements of a secondary-line RPA case.  Why the third condition?  The more of a seller’s product that is channeled through intermediate distributors, the less able an RPA defendant, whether a manufacturer (e.g. Coca-Cola) or a retailer (e.g. Walmart), is able to use the cost justification defense to argue it is the cost of distribution that justifies different prices. 

Recall that the cost justification defense allows an RPA defendant to appeal to differences in “the cost of manufacture, sale, or delivery resulting from the differing methods or quantities in which such commodities are to such purchasers sold or delivered.”  When independent distributors carry the product, the cost differential in serving one customer or another is very likely to be the same.  

For those keeping score at home wondering why on earth the FTC would begin its RPA crusade with a lawsuit against Southern Glazer, a liquor distributor, this is at least part of your answer.  Beer, wine, and spirits are sold through a complex system of state and federal regulation involving a mandatory “three-tier system” in nearly all of the United States.  The “three-tier” system mandates that the distiller / brewery/ winery cannot vertically integrate intro distribution; the distributor cannot vertically integrate into production or retail; and the retailer cannot distribute or produce.  Each layer of the vertical supply chain is separate.  It is a different topic for another day, but state regulation of alcohol, including but not limited to the three-tier system, is wildly anticompetitive.  I’ll write some about this later, but if you’re very interested start here and then read this fun (but I think very wrong) article from Tyler Cowen defending alcohol monopolies.  

What does the three-tier system buy the RPA plaintiff?  Plenty.  The distributor only distributes.  Southern Glazer sells to Costco and other retailers and in each case the distribution cost is likely the same.  No doubt Southern Glazer will argue the costs of selling to Costco and a supermarket or a convenience store are quite different – but the costs are observable.  One reason Costco is such an efficient retailer is that it strategically vertically integrates into parts of the supply chain in order to keep costs down.  Perhaps the best known example of this strategy is deploying vertical integration to keep the costs and prices of its famous hot dog combo down (not to mention rotisserie chicken): 

Roland M. Vachris, Costco’s chief executive officer, is well aware of the company’s foothold among consumers in search of the retailer’s food court offerings. Vachris said during the company’s earnings call in May that the retailer has been able to maintain the its hotdog and coke at $1.50 because of “vertical integration and sourcing as the need arises.”  Vachris said that in order for Costco to maintain the hotdog price combo at that level it has had to open its own meat plants. “We had to get involved because we didn’t have a partner that was willing to expand into that area.” Costco took that same approach with optical lenses and chicken, he said.

The law prevents vertical integration by Costco (or Jack Daniels) into alcohol distribution.  That simplifies life for the FTC as a plaintiff in a secondary-line RPA case against Southern Glazer or a 2(f) case against Costco directly.  Off the table are arguments that, when Southern Glazer sells to Costco (or Walmart or Target) at a much lower cost of distribution than sales to other retailers it is because of vertical integration or strategic location of distribution centers.  Of course, Southern Glazer can and will still point to some cost differences – large retail customers may be less costly to serve for other reasons too!  

But when the retailer can vertically integrate into distribution, think about the advantages for both potential targets of an RPA suit: the seller and the large, vertically integrated retailer. The vertically integrated retailer is better off because it changes the nature of the transaction.  It is now simultaneously buying the product from the seller, but is itself also selling distribution services.  Whether it negotiates a lower wholesale price from the seller or extracts some other concessions, it is difficult for an RPA plaintiff to argue that the costs of serving the vertically integrated and independent retailers are similar.  Indeed, it will be difficult for the RPA plaintiff in some such cases to persuade a court the sales to integrated retailers (product + distribution) and independents (just the product or just distribution) are of “like grade and quality,” which the RPA requires. 

The bottom line is that sellers, distributors, and retailers face potential RPA liability because they sell at different prices to vertically integrated and non-vertically integrated retailers.  As RPA enforcement becomes sufficiently widespread, the incentives become more powerful to adapt by reorganizing economic activity exclusively within larger, vertically integrated players.  Costco, Walmart, and Target have the incentives to vertically integrate into distribution more broadly so that they are not punished for their greater efficiency.  Sellers have powerful incentives to sell more volume to large, vertically integrated players rather than independents because it is different prices offered to competing players that generates RPA liability.  Why bother with the independents? 

By no means am I arguing that a handful of still-to-come RPA cases – presuming the FTC does file them – will immediately impact vertical integration in any meaningful sense.  But where there are social costs imposed by punishing efficient retailers because distributors sell both to them and independents, there are gains to retailers from vertical integration into distribution and gains to suppliers of selling exclusively to large retailers rather than independents. 

The other social cost of RPA enforcement is in the form of making American consumers poorer by raising prices as a first order effect.  But I would not be surprised if we also see more attempts to re-organize economic activity to mitigate the effects of RPA enforcement if it becomes common enough.  Still, I’m guessing that raising barriers to entry for small players and independents by providing incentives to vertically integrate was not what the Neo-Brandeisians had in mind.  

I hope you enjoyed our three part series on the Robinson Patman Act.  Here are links to Part I and Part II in case you missed them.  And no doubt we will return to the subject when and if the FTC files its complaints.  

Solving the Robinson-Patman Act / Consumer Welfare Conundrum

The Robinson-Patman Act Meets Economics

Later this week, from Robinson-Patman to the Rule of Law.  Next newsletter we will discuss the FTC’s Exxon-Pioneer Consent Decree and the Rule of Law.