Like Thom, I also have spent the last few weeks reading Herbert Hovenkamp’s excellent new antitrust book, The Antitrust Enterprise: Principles and Execution. I am looking forward to Thom’s review in the Texas Law Review, and wholeheartedly agree with him that Hovenkamp’s book is an important and significant contribution to the antitrust literature (see also Randy Picker’s book review here describing “The Antitrust Enterprise as The Antitrust Paradox for a post-Chicago antitrust landscape”). I’m still digesting most of the book, and perhaps will share some more thoughts in this space later on, but thought I would chime in with some thoughts on two issues relevant to my own research on slotting contracts, discounts, and competition for product distribution.
Hovenkamp endorses a generally sensible approach to antitrust treatment of manufacturer payments, e.g. quantity and market-share discounts, slotting allowances, and Lepage’s-type bundled discounts. Hovenkamp recognizes that discounting is a “pervasive feature of the American economy,” and that “quantity and market-share discounts are virtually always competitive unless they amount to outright exclusive dealing,” but he adds that “even exclusive dealing is competitively harmless in most circumstances.” Hovenkamp appears to have greater reservation about the potentially exclusionary effects of bundled discounts, but ultimately concludes that administrative costs justify a lenient antitrust rule:
Even though the theory of the bundled discount is properly analogized to tying or exclusive dealing rather than predatory pricing, an administratively prudent rule might insist on a showing the the discounted package is priced below average variable cost.
As I’ve noted in this space previously, and this paper (now in print at 23 Yale Journal on Regulation 169) antitrust rules should reflect the welfare benefits generated as shelf space payments are ultimately passed on to consumers:
If the retail sector is competitive, which is almost always the case as a result of low barriers to entry, these payments are passed on to consumers regardless of form. These payments create first order benefits for consumers in the form of lower prices and higher quality. A coherent antitrust policy will recognize that these payments are a form of the competitive process, namely price competition, and should be treated as such.
Further, where anticompetitive exclusion is the competitive concern, antitrust law would be best served by establishing safe-harbors for distribution contracts unlikely to create anticompetitive effect, i.e. short-term contracts or contracts foreclosing less than 40% of distribution assets. This approach applies to competition for distribution generally, and is not limited to bundled discounts. Thom’s post and analysis in his Minnesota Law Review piece offer sensible and similarly-minded policy proposals for evaluating bundled discounts. With all of that said about the general sensibility of Hovenkamp’s approach here, I have two quibbles upon which I will expand below the fold.
The first is a tension in Hovenkamp’s analysis of bundled discounts and exclusive dealing. Hovenkamp concludes that adminstrative costs justify a predatory pricing-type rule in the context of for bundled discounts where the anticompetitive mechanism is de facto “foreclosure” or deprivation from distribution resources (i.e. shelf space) that would prevent rivals from achieving minimum efficient scale and extend the duration of monopoly by increasing barriers to entry. One would think that it would follow from Hovenkamp’s position that a predatory pricing-type rule would also be sensible for exclusive dealing and tying arrangements where the anticompetitive mechanism is the economic equivalent. To the contrary, Hovenkamp advocates rule of reason analysis (p. 201) for exclusive dealing and tying, noting that “foreclosure concerns can be assessed meaningfully only via the rule of reason” and that “the antitrust law of exclusive dealing,” which generally requires proof of substantial foreclosure as a necessary condition of competitive harm, “seems to be on the right track.”
I should note that I agree with Hovenkamp’s conclusion that all exclusionary practices should be assessed under the rule of reason because they are likely to be pro-competitive. In addition to the pass-through of payments to retailers already discussed, exclusivity terms have a number of well-known efficiency justifications and are common in retail distribution. There are, of course, reasonable ways to reconcile the apparent tension between Hovenkamp’s positions on bundled discounts (as de facto exclusives) and exclusive dealing contracts, i.e. that the administrative costs of assessing foreclosure and exclusion with bundled discounts are considerably greater than the rule of reason analysis for exclusive dealing. But Hovenkamp does not explicitly make this claim. Nor would I find such an argument persuasive. The standards for above-cost predation should require proof of substantial foreclosure and anticompetitive effect for the reasons that Hovenkamp cites, but it does not make economic sense to have two different standards for these claims based on the same mechanism of exclusion.
My second quibble is to do with Hovenkamp’s treatment of slotting allowances, manufacturer payments to retailers for shelf space. First, Hovenkamp should be applauded for addressing shelf space monopolization issues in such great detail in his book (including the excellent discussion of the Conwood case, which Hovenkamp correctly describes as an anticompetitive decision). Understanding the economic function of various forms of shelf space contracts and distribution arrangements has been a central concern of the antitrust community for the past several years, as evidenced by the substantial growth in private and public litigation, the FTC Workshop and Study on slotting allowances and category management, and the outright ban of some of these practices in other jurisdictions. With this in mind, Hovenkamp correctly recognizes that:
One poorly understood discount that is sometimes challenged under the antitrust laws is “slotting” fees, or upfront fixed fees by which manufacturers purchase scarce shelf space . . .. It is difficult to make a cogent argument why slotting fees are anticompetitive, with the possible exception of those resulting in a below-cost price.
The main function of slotting fees is to transfer risk from the retailer to the manufacturer . . .. The slotting fee shows the merchant that the manufacturer’s promise of good sales is more than empty words. The manufacturer who pays a $100 slotting fee is betting that the sales of its product in that store will be sufficiently high that the net discount that results from the fee will be affordably small. Significantly, any equally efficient and confident firm could match the fee. A firm that claims it cannot afford such a fee is in fact telegraphing that it lacks confidence that sales of its product will be high enough to merit the grocer’s investment.
While I agree that it is difficult to come up with a persuasive economic argument that slotting is anticompetitive, theories are abound in the antitrust literature: the FTC Workshop, FTC v. McCormick, various game theoretic models in the economics literature, etc. The reason that these theories are not persuasive is not internal failure of logic, but because anticompetitive theories both fail to explain and are contradicted by available data. Unfortunately, this is also true of the pro-competitive explanation Hovenkamp attributes to slotting: product risk.
Product risk does not explain slotting fees. As Ben Klein and I explain in The Economics of Slotting Contracts (forthcoming in JLE), the “risk” explanations for slotting suffer from empirical and theoretical defects. Empirically, the notion that slotting “signals” the likely success of a new product does not explain the use of slotting on established products such as Coca-Cola or Marlboro cigarettes which do not involve such product risk. Nonetheless, Klein and Wright provide evidence that slotting frequently occurs with incumbent products. My own empirical evaluation of the competitive effects of slotting contracts in military commissaries, Slotting Contracts and Consumer Welfare (forthcoming in Antitrust Law Journal) utliizes a set of slotting contracts consistent of entirely incumbent products. Further, the signaling theories of slotting do not address why manufacturers do not resort to alternative contractual arrangements such as introductory price allowances based on sales and liberal return policies to insure against this risk (and therefore are not capable of explaining the slotting data, e.g., why slotting has increased over time, why these contracts exploded in popularity in the mid 1980s, variance in payments across markets, etc.). As Klein and Wright note, perhaps most importantly, the signaling models are flawed because:
The transactions cost and risk cost theories do not attempt to answer the fundamental economic question underlying the existence of slotting contracts, namely why consumers do not pay for the higher costs of supermarket operations in a higher retail price, rather than having manufacturers cover the increased costs with a per unit time slotting fee.
I do not wish to suggest here that Hovenkamp’s analysis of shelf space monopolization via slotting, discounts, and category management is wholly misguided. It is not. Hovenkamp understands that competition for distribution is generally pro-competitive and argues that administrable antitrust rules should guide the competitive process. This is a sensible approach and all of this is to the good. Ultimately, I believe that Hovenkamp’s policy conclusions about shelf space payments and discounts are helpful (even if, in the case of slotting, for the wrong reason), but am less persuaded that bundled discounts and exclusive dealing should receive different treatment, especially since rational antitrust plaintiffs can be expectedly to substitute towards the more friendly standard. Hovenkamp’s book offers a comprehensive treatment of these issues, and it will surely generate a good deal of fruitful discussion aimed at the design of administrable standards capable of identifying anticompetitive conduct without overdeterring consumer welfare-enhancing competition for distribution.