Bargaining Over Loyalty

id=1937658; Einer Elhauge & Abraham L. Wickelgren, Robust Exclusion Through Loyalty Discounts (Harvard Law Sch. Pub. Law & Legal Theory Working Paper Series, Paper No. 10-15, 2010) [hereinafter Elhauge & Wickelgren, Robust Exclusion], available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1544008; Patrick Greenlee & David Reitman, Competing with Loyalty Discounts (Econ. Analysis Grp., Discussion Paper No. 04-2, 2005); Leslie M. Marx & Greg Shaffer, Rent Shifting, Exclusion, and Market-Share Discounts (June 2004) (unpublished manuscript), available at http://www.simon.rochester.edu/fac/shaffer/ Published/rentshift.pdf; Janusz Ordover & Greg Shaffer, Exclusionary Discounts (Ctr. for Competition Policy, Working Paper No. 07-13, 2007), available at http://papers.ssrn.com/sol3/ papers.cfm?abstract_id=995426. 15. See infra text accompanying notes 230–40. 16. See infra text accompanying notes 246–56. CRANE.FINAL.RESUBMIT.OC (DO NOT DELETE) 12/18/2013 9:37 AM 2013] Bargaining over Loyalty 257 European Union, and discusses the way that U.S. antitrust law on loyalty discounts is likely to evolve in light of recent precedent—not primarily by the development of new legal rules but by the expression of judicial maxims such as the quotation from the Virgin Atlantic decision at the beginning of this Article. This Article’s second major contribution is to answer two developing criticisms of loyalty discounts that have the potential to turn into antiloyalty judicial maxims. The first of these is that loyalty discounts need not be— and often are not—true discounts, but rather disloyalty penalties. This claim is economically implausible since it would have the seller giving the buyer a choice of accepting either a price above the profit-maximizing monopoly level or else an onerous contractual term, which would be akin to a price above the profit-maximizing monopoly level. Either scenario would effectively cause the monopolist to exceed the profit-maximizing monopoly price and hence be unprofitable. The second criticism of loyalty discounts is that they soften competition between sellers—essentially that they facilitate supracompetitive seller pricing even without excluding any seller from the market. Part II shows that the assumptions underlying this claim are restrictive and not generalizable. This Article’s final major contribution, made in Part III, is to reorient the conversation away from an assumption that loyalty incentives are sellerinitiated strategies. Rather, the available evidence suggests that loyalty incentives are often bargaining chips in negotiations between sellers and buyers—invoked by customers as often as suppliers in return for other concessions. Thinking about loyalty incentives as bargaining chips does not dispel the possibility that such provisions can have exclusionary effects, but it does suggest that courts should be cautious about discouraging the use of loyalty incentives, which may take away a chip that buyers could otherwise invoke to improve their position. I. Foundational Considerations A. Loyalty and Volume Loyalty provisions come in a variety of forms. The strongest form is a pure exclusive dealing agreement in which the buyer promises to buy all of its requirements from the supplier and not to purchase from any other supplier. Short of this, contracts sometimes contain partial exclusive dealing clauses that commit the buyer to make a specified level or 17. Virgin Atl. Airways v. British Airways Plc, 257 F.3d 256 (2d Cir. 2001). 18. See generally 11 HERBERT HOVENKAMP, ANTITRUST LAW: AN ANALYSIS OF ANTITRUST PRINCIPLES AND THEIR APPLICATION ¶ 1821a2 (3d ed. 2011) (contrasting the stringent nature of exclusive dealing with market share discounts and market share dealing requirements). CRANE.FINAL.RESUBMIT.OC (DO NOT DELETE) 12/18/2013 9:37 AM 258 Texas Law Review [Vol. 92:253 percentage of its purchases from the seller. Often sellers seek to induce loyalty rather than to require it. Loyalty inducement provisions also take a good many forms, but their common denominator is an option on the buyer’s part to secure a better price by demonstrating greater loyalty. Two common forms are market share discounts and bundled discounts. Bundled discounts offer a buyer a better price for purchasing minimum amounts of the seller’s product across two or more separate product lines. For example, in one of the leading recent bundled discount cases, the conglomerate manufacturer 3M offered retailers rebates that were conditioned on the retailer making minimum purchases on six of 3M’s product lines, including Health Care Products, Home Care Products, Home Improvement Products, Stationery Products, Retail Auto Products, and Leisure Time. Unlike a single-product volume discount, the customer can only achieve bundled discounts or rebates by demonstrating loyalty in a number of separate buckets of purchases. Although bundled discounts partake of many of the attributes of single-product loyalty discounts, they add significant complexities. Bundled discounts create different kinds of exclusionary effects— particularly the potential to exclude rivals that do not sell the dominant firm’s full product line. They also may exhibit different sorts of efficiencies or procompetitive justifications—such as the potential to eliminate double marginalization—that would not generally be true of single-product loyalty discounts. Further, bundled discounts raise unique theoretical questions about the plausibility of a dominant firm’s exclusionary strategy—such as whether it would be rational for a firm to use a bundled discount to leverage market power in one market to obtain a monopoly in a second market in light of the fact that raising the price in the second market might reduce sales in the first market if the two goods are 19. See Tom et al., supra note 14, at 621–22. 20. See Daniel A. Crane, Mixed Bundling, Profit Sacrifice, and Consumer Welfare, 55 EMORY L.J. 423, 425 (2006) (introducing the economics of bundled discounting). See generally Einer Elhauge, Tying, Bundled Discounts, and the Death of the Single Monopoly Profit Theory, 123 HARV. L. REV. 397 (2009) (discussing the exclusionary potential of bundled discounts); Herbert Hovenkamp & Erik Hovenkamp, Complex Bundled Discounts and Antitrust Policy, 57 BUFF. L. REV. 1227 (2009) (analyzing a variety of bundling practices and concluding that some should be considered anticompetitive); Thomas A. Lambert, Appropriate Liability Rules for Tying and Bundled Discounting, 72 OHIO ST. L.J. 909 (2011) (responding to and critiquing Elhauge’s proposed liability rule and arguments regarding the effects of tying). 21. LePage’s Inc. v. 3M, 324 F.3d 141, 154 (3d Cir. 2003) (en banc). 22. See Crane, supra note 20, at 443–47. 23. Id. at 434–36; Erik Hovenkamp & Herbert Hovenkamp, Tying Arrangements and Antitrust Harm, 52 ARIZ. L. REV. 925, 958–61 (2010). 24. But see Sreya Kolay et al., All-Units Discounts in Retail Contracts, 13 J. ECON. & MGMT. STRATEGY 429, 434–35 (2004) (discussing the potential of even single-product market share discounts to eliminate double marginalization). CRANE.FINAL.RESUBMIT.OC (DO NOT DELETE) 12/18/2013 9:37 AM 2013] Bargaining over Loyalty 259 complements. Because of these significant distinctions, bundled discounts merit separate consideration from single-product market share or other loyalty discounts and are beyond the scope of this Article. Market share discounts are the paradigmatic single-product loyalty incentive. They operate by granting the buyer a better price if it purchases specified percentages of its requirements from the seller. Market share discounts are sometimes graduated—for example, a buyer receives a 5% discount for purchasing 60% or more of its requirements from the seller, a 7% discount for purchasing 75% or more, and a 9% discount for purchasing 90% or more. Also, market share discounts may apply only to incremental dollars (i.e., to all purchases above 60%) or retroactively to the first dollar. A loyalty discount can be given instantaneously at the point of sale or rebated at some later time, such as at year’s end. How loyalty discounts are structured is often significant in determining whether they can have exclusionary effects. For example, first-dollar rebates are usually considered more problematic than incremental-dollar discounts, since smaller rivals of the seller have to compete against price concessions given across a far greater swath of sales. Contracts with claw-back features, where the seller grants the buyer 25. See Elhauge, supra note 20, at 403–19. The Chicago School of economic analysis argued that it would be irrational for a firm with a monopoly in market A to attempt to leverage its power into a complementary market B, since increasing the price of one product leads to a diminution in the demand for its complements. Hence, by leveraging monopoly power and attempting to extract a second monopoly profit, the dominant firm would simply be cannibalizing its own profits in the leveraging market. See ROBERT H. BORK, THE ANTITRUST PARADOX: A POLICY AT WAR WITH ITSELF 372–75, 380–81 (1978) (arguing that there is no reason for a firm with one monopoly to try to use tying to extract a second monopoly profit); RICHARD A. POSNER, ANTITRUST LAW 197–99 (2d ed. 2001) (contending that tying will not allow the firm to gain a second monopoly in the case of either complementary or unrelated goods); Ward S. Bowman, Jr., Tying Arrangements and the Leverage Problem, 67 YALE L.J. 19, 20–23 (1957) (concluding that a firm is unable to extract a second monopoly profit because any price increase to one good must be offset by a corresponding decrease in the other good); Aaron Director & Edward H. Levi, Law and the Future: Trade Regulation, 51 NW. U. L. REV. 281, 290–92 (1956) (proposing that any attempt to impose coercive restrictions on customers will be successful only if the price that would be charged without the restriction is reduced). Elhauge argues that the one-monopoly-profit theory overlooked a number of ways in which leverage could be profitable. Elhauge, supra note 20, at 403–19. 26. See Herbert Hovenkamp, The Obama Administration and Section 2 of the Sherman Act, 90 B.U. L. REV. 1611, 1649–50 (2010). A variation on a market share requirement is a retailshelf-placement requirement. In Church & Dwight Co. v. Mayer Labs., Inc., 868 F. Supp. 2d 876, 887 (N.D. Cal.), vacated in part, No. C-10-4429 EMC, 2012 WL 1745592 (N.D. Cal. May 16, 2012), for example, a condom manufacturer granted retailers different levels of rebates for maintaining its products in various percentages of the retailer’s shelf space dedicated to condoms: “[A] 55% tier (awarding a 4.0% rebate for 55% or more of a retail chain’s display space), a 65% tier (awarding a 7% rebate for 65% or more of the display space), and a 70% tier (awarding a 7.5% rebate for 70% or more of the display space).” 27. See Jacobson, supra note 14, at 1–2; Duncan & McCormac, supra note 14, at 134. 28. See Jonathan M. Jacobson, Towards a Consistent Antitrust Policy for Unilateral Conduct, ANTITRUST SOURCE, Feb. 2009, art. 3, at 1, 6, available at http://www.americanbar.org/content/ CRANE.FINAL.RESUBMIT.OC (DO NOT DELETE) 12/18/2013 9:37 AM 260 Texas Law Review [Vol. 92:253 a favorable price on the assumption that it will meet a loyalty threshold, subject to a repayment obligation in the event the buyer does not meet the threshold, may create particular antitrust risks insofar as buyers may be loath to run the risk of incurring large lump-sum penalties at year’s end and hence remain strictly loyal to the seller. One of the frequently discussed questions with respect to market share discounts is why a seller who wants to reward high-volume customers should not simply offer a traditional volume discount. Before getting to some of the answers, it is worth noting that sometimes volume and loyalty discounting is substantively equivalent. Volume discounts and loyalty discounts can be identical in operation. Suppose, for example, that the buyer has a stable need for one hundred tons of salt per year. If the seller offers the buyer a 5% price reduction for buying eighty tons of salt or 80% of its salt requirements per year, the effect on the buyer’s incentives will be identical assuming its buying needs stay constant. On the other hand, market share discounts often differ from volume discounts in significant ways. In several circumstances, market share discounts may be more advantageous to the buyer than volume discounts. First, market share discounts have the effect of shifting risks of changing market circumstances from buyers to sellers in ways that volume dam/aba/publishing/antitrust_source/Feb09_SourceFull2_26f.authcheckdam.pdf (explaining that “‘first dollar’ discounts[] may provide especially strong inducements—in some instances, outright coercion—because they apply not only to the contested volume but to all of the customer’s purchases, enhancing a loss if the percentage commitment is not fulfilled”); Robert H. Lande, Should Predatory Pricing Rules Immunize Exclusionary Discounts?, 2006 UTAH L. REV. 863, 863–64 (“Unlike ‘regular’ discounts, which are almost always procompetitive, retroactive discounts have a strong exclusionary and anticompetitive potential.”). 29. See Jacobson, supra note 28. Allegations about claw-back provisions have been at issue in some recent loyalty discount cases. See, e.g., Se. Mo. Hosp. v. C.R. Bard, Inc., 642 F.3d 608, 617 (8th Cir. 2011). 30. See, e.g., Evaluating the Competitive Effects of Exclusive Dealing Agreements, ANTITRUST SOURCE, Nov. 2005, art. 2, at 1, 6, available at http://www.americanbar.org/content/ dam/aba/publishing/antitrust_source/Nov05_FullSource11_29.authcheckdam.pdf (“[W]hy reward your best customers with a ‘loyalty’ discount? Why not do it instead through a less restrictive alternative like a volume discount?”). On the distinction between market share and volume discounts, see ZF Meritor, LLC v. Eaton Corp., 696 F.3d 254, 265 n.6 (3d Cir. 2012) (citing PHILLIP E. AREEDA & HERBERT HOVENKAMP, 3B ANTITRUST LAW: AN ANALYSIS OF ANTITRUST PRINCIPLES AND THEIR APPLICATION ¶ 768b4 (3d ed. 2008)). 31. A case in point is then-Judge Breyer’s decision in Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227 (1st Cir. 1983), which involved the market for snubbers, safety devices used in nuclear power plants. The defendant offered a major customer a large discount if it would agree to purchase large quantities of snubbers, amounting to a large share of its expected purchases, over a two-year period. Id. at 228–29. Since the customer’s snubber needs were stable and predictable, it probably would have made little difference if the supplier made the discount contingent on loyalty or volume. See id. at 229 (noting that the customer was able to predict its needs two years in advance and that its snubber needs were identical for both years). CRANE.FINAL.RESUBMIT.OC (DO NOT DELETE) 12/18/2013 9:37 AM 2013] Bargaining over Loyalty 261 discounts do not. For example, Herbert Hovenkamp has explained Intel’s market share discounts as a means of shifting the risk of a weakening market from original equipment manufacturers (OEMs) like Dell and HP to Intel. If the computer market weakens more than expected, Dell and HP might not be able to meet a contractually specified volume threshold and hence might lose a volume-based discount. However, if to obtain Intel’s lowest price they must just buy a specified percentage of their central processing unit needs from Intel—say 80%—they can continue to claim the best price even in a weak computer market. Second, and in the same vein, market share discounts may be used to guarantee the supplier a minimum volume of sales when the requirements of a group of customers are unpredictable. To stay with the computer industry, suppose that Intel will be able to optimize its planning and achieve economies of scale if it knows that it will sell at least one million central processing units (CPUs) in the coming year. Although it makes a fairly strong prediction that the total volume of CPU sales in the market will be around two million, the OEMs are engaged in a fierce market share battle of their own, and the CPU requirements of any individual OEM are hard to determine given the vagaries of the market. Intel may identify a group of OEMs that are likely to purchase around 1.25 million collectively, although the distribution of purchases within the group is uncertain. If each of the OEMs in this group agrees to purchase 80% of its requirements from Intel, Intel can count on making a million CPU sales in the coming year from this group of customers. From Intel’s perspective, it is beneficial to offer a discount in exchange for a market share commitment so that Intel can plan on the level of sales it will make in the coming year and perhaps optimize its production facilities. From the OEMs’ perspective, the deal is also beneficial. The OEMs secure a more favorable price and one that does not 32. Herbert Hovenkamp, The Federal Trade Commission and the Sherman Act, 62 FLA. L. REV. 871, 889 (2010). Hovenkamp explained that a seller like Intel uses market share discounts rather than volume discounts in order to shift the costs of market downturns from its customers to itself, writing: A quantity discount attaches to a specified number of chips, and if the market becomes weak and the computer maker’s sales fall below that number, the computer maker must pay the higher price. By contrast, a market share discount attaches to, say, 90% of the buyer’s sales, whatever they happen to be. So the market share discount offers the computer maker the lower price, even if the market becomes weak, provided that the computer maker purchases its requisite percentage of chips from the seller.