This document is available in three formats: this web page (for browsing content), PDF (comparable to original document formatting), and Word. To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official signed copy, please contact the Antitrust Documents Group.

Preliminary Draft: 11/10/06
Do not cite or distribute without permission



Benjamin Klein


Joshua D. Wright*


Category management is a business practice whereby a retailer designates a manufacturer as a category manager or captain and gives the designated manufacturer authority concerning retail shelf space allocation within a product category. In return for shifting brand stocking decisions [as well as promotion, product assortment, and inventory decisions] to the designated manufacturer, the retailer receives a lower wholesale price or a per unit time payment from the manufacturer. This paper analyzes the antitrust law and economics of category management contracts, demonstrating that they are an element of the normal competitive process that benefits consumers and challenging the increased antitrust scrutiny that has been applied to such arrangements as exemplified by the Sixth Circuit's recent decision in Conwood Co. v. United States Tobacco Co. We show that the economics of category management contracts is not fundamentally different from exclusive shelf space contracts-- control over the shelf space allocation decision is merely shifted from the retailer to a manufacturer with the manufacturer becoming the transactor that can violate the implicit contract and the retailer becoming the policer of the contract. Mistaken reasoning regarding fiduciary obligations and horizontal versus vertical contracts has led to the placement of greater antitrust scrutiny on category managers than on firms that have negotiated more restrictive fully exclusive distribution contracts.


  1. Introduction
  2. Manufacturer Contracting for Exclusive Retail Shelf Space
    1. Manufacturer Competition for Shelf Space
    2. Exclusive Shelf Space Contracts

  3. Category Management Contracts
  4. Antitrust Analysis of Category Management
    1. Category Management in the Courts: Conwood Co. v. United States Tobacco Co.
    2. There Was Insufficient Evidence to Conclude that UST's Conduct Was Anticompetitive

  5. Conclusion

  1. Introduction

Distribution contracts that explicitly or implicitly restrict the display, promotion, or sale of rival products by retailers have recently been the subject of antitrust litigation(1) and a Federal Trade Commission study.(2) In the multi-product retailer setting, shelving restrictions placed on rival products come in a variety of forms ranging from exclusive dealing arrangements, whereby the retailer may not stock any rival brands, to limited exclusives, which grant a manufacturer access to preferred shelf space but do not completely exclude all rivals.(3) Category management contracts, which shift control of the retailer's shelf space within a product category to a manufacturer, are another form of limited exclusive, where the manufacturer determines the quantities of other, highly demanded products to be stocked.

Klein and Wright demonstrate that retailer shelf space slotting arrangements solve a pervasive incentive incompatibility between manufacturers and retailers over the supply of promotional shelf space, which induces some consumers to purchase the displayed product who would not otherwise do so, but does not induce consumers to shift between stores.(4) Because there are little or no inter-retailer competitive effects from the supply of promotional shelf space, retailers do not have the incentive to provide the joint profit-maximizing quantity of promotional shelf space and manufacturers must contract for such space. While this incentive incompatibility explains the existence of shelf space contracts, it does not explain why shelf space arrangements sometimes include restrictive shelving terms on rival manufacturers. Klein and Murphy demonstrate that exclusive or partially exclusive shelf space arrangements allow a retailer to obtain a greater rate of return on its shelf space by essentially promising to transfer all its customers' purchases to the selected manufacturer, and therefore increasing the elasticity of demand manufacturers face in bidding for the retailer's shelf space.(5) This increase in the value the retailer receives for shelf space is passed on to consumers where inter-retailer competition is significant and, therefore, likely to benefit consumers. Our focus is on category management as a specific form of limited exclusive distribution arrangement and the conditions under which shelf space allocation of rival brands is sometimes shifted to the manufacturer.

For example, consider the shelf space contracts negotiated by Gruma Corporation, the nation's largest tortilla manufacturer, who was alleged by a number of regional manufacturing rivals to have foreclosed competitors from effective retail distribution.(6) The Gruma shelf space contracts with supermarkets allowed Gruma, in return for slotting fees and other financial incentives, to "control the placement, location, availability, visibility and promotional activity of competing retail tortillas."(7) This paper provides a pro-competitive economic explanation for category management retail shelf space arrangements.

Section II summarizes the economics of manufacturer purchase of exclusive shelf space contracts, identifying conditions under which such contracts are an efficient element of the competitive distribution process. We show how exclusivity is likely to be a consequence of competition for shelf space when a retailer's consumers are "loyal," in the sense that in response to exclusivity in a particular product category most consumers will not switch to retailers that stock alternative brands. In these circumstances a retailer can obtain a higher slotting fee with the offer of an exclusive that, by restricting the ability of individual consumers to exercise their brand preferences, internalizes each consumer's independent buying decision and increases the elasticity of demand faced by each supplier. Exclusivity, therefore, may increase the value the retailer receives for its shelf space and, when overall inter-retailer competition is significant, this value will be passed on to consumers. Retailers are using exclusivity to act as efficient bargaining agents for their consumers.

Section III then analyzes the more common case where some consumers' demand for a particular brand is high and, therefore, where the efficient shelf space contract involves some form of limited exclusive. We show that category management contracts are analytically similar to limited exclusive shelf space contracts where control over the shelf space allocation decision is shifted from the retailer to a manufacturer. SectionIV then turns to the role of antitrust in regulating exclusive dealing and category management relationships, analyzing Conwood through the lens of our model of category management and competition for distribution and concluding that it perversely imposes a greater fiduciary obligation for category managers than monopolists implementing exclusive shelf space contracts.

  1. Manufacturer Contracting for Exclusive Retail Shelf Space

We analyze the economic forces that lead to exclusive shelf space arrangements in two steps. We first demonstrate that where there are little or no inter-retailer competitive effects from the supply of promotional shelf space, retailers do not have the incentive to provide the joint profit-maximizing quantity of promotional shelf space, and manufacturers must contract for such space rather than relying on retailers to independently choose which products to stock and prominently display.(9) We then turn to the question of when manufacturers and retailers will adopt exclusive shelf arrangements which explicitly limit the retailer's ability to promote rival products.

    1. Manufacturer Competition for Shelf Space

Retailer shelf space is a form of promotion because it induces some consumers to purchase the particular displayed product who would not otherwise do so.(9) Unlike abstract economic models where the sole function of retailing is to reduce shopping costs, retailers have the ability to influence consumer purchases with stocking and display decisions. The reservation value of marginal consumers who would not otherwise purchase the product increases in response to prominent shelf display resulting in greater sales. It is useful to think of promotional shelf space as a targeted effective price discount to these consumers and manufacturers competing for the incremental sales.

Since consumers are not willing to pay for promotional shelf space, but it induces profitable incremental sales for manufacturers, manufacturers will generally want greater retailer shelf space than retailers are willing to supply on their own. Retailers deciding how much promotional shelf space to supply will not take account of the manufacturer's profit margin on the incremental sales produced by the promotional shelf space. This incentive incompatibility should therefore be especially significant when the manufacturer supplies differentiated products where the wholesale price is much greater than its marginal cost of production.(10)

It is important to recognize that this incentive incompatibility does not imply that retailers will always supply less than the desired amount of all forms of non-price competition. For instance, retailers have the incentive to supply forms of non-price amenities, such as free parking, whenever a sufficient number of marginal consumers value the service. In this case, inter-retailer competition forces retailers to provide the non-price service. However, it is reasonable to assume that a retailer's decision to prominently display a particular brand does not generate significant inter-retailer effects in the form of consumers switching to competing retailers because their desired brand is not prominently displayed. Therefore, retailers will supply less than the jointly profitable amount of promotional shelf space.

This competitive distortion is not a significant problem with respect to retail pricing decisions because a lower retail price has a much larger impact on the retailer's individual demand than the manufacturer's demand because a reduction in retail prices involves significant inter-retailer demand effects as well as inter-manufacturer demand effects for the retailer, whereas the manufacturer only experiences the latter. While retailers also do not take into account the manufacturer's much larger profit margin on incremental sales when lowering prices, the manufacturer can be assured that retail price competition will be approximately optimal because the significant inter-retailer demand effects largely offset the manufacturer's larger margin, nearly eliminating this distortion.

Because retailers will often have an insufficient incentive to supply the jointly profit maximizing level of promotion, manufacturers must compensate the retailer in order to assure the supply of increased retailer promotion. Competitive manufacturer bidding for shelf space will determine which brands will be prominently displayed by retailers as well as the equilibrium value of promotional shelf space. Whether the manufacturer payments are per unit time (e.g., slotting fees) or per unit sale (a wholesale price discount), the payments compensate retailers for "extra" shelf space. It is therefore useful to think of manufacturer competition for shelf space as an efficient element of the competitive contracting process designed to ensure the provision of the jointly profit-maximizing level of shelf space.(11) Because competitive shelf space payments can be expected to be passed on to consumers, this competitive process benefits consumers in the form of increased non-price amenities and lower prices.(12)

    1. Exclusive Shelf Space Contracts

While we have demonstrated that manufacturers contracting for shelf space solves a pervasive incentive incompatibility between manufacturers and retailers over the supply of promotional shelf space, that analysis does not explain why shelf space arrangements sometimes include shelving terms explicitly or implicitly restricting rivals. These full or limited exclusive retail contracts have recently been the subject of antitrust scrutiny on the theory that various forms of shelving restrictions enable dominant manufacturers to exclude rivals and harm consumers.(13)

Klein and Murphy provide a procompetitive explanation for exclusive or partially exclusive shelf space arrangements.(14) Specifically, exclusivity allows the retailer to obtain a greater rate of return on its shelf space by committing to transfer all or a significant portion of its' customers purchases to the selected manufacturer, therefore increasing the elasticity of demand faced by manufacturers bidding for shelf space. Retailers effectively act as bargaining agents for their consumers by internalizing what would otherwise be each consumer's independent buying decision if multiple products were on the shelf. Furthermore, exclusive dealing may be used to facilitate contracting over promotional shelf space by efficiently defining what the manufacturer is buying.

When retail display decisions do not induce significant inter-retailer effects, it is important to recognize that the reduction in variety associated with full or limited exclusives is more than offset by the increase in payments for shelf space. The increased payments associated with exclusive shelf space arrangements improve consumer welfare because rigorous competition at the retail level ensures that these payments are ultimately passed on to customers in the form of lower prices or increased quality.

Exclusive shelf space contracts are not explained by the possibility of anticompetitive effects, which has been the primary competitive concern expressed by antitrust agencies.(15) It is well known that slotting contracts and other exclusive distribution arrangements may produce anticompetitive effects if and only if a dominant manufacturer can control a sufficient amount of distribution for a sufficient period of time, so that rivals are effectively prevented from reaching minimum efficient scale. However, this type of anticompetitive mechanism is functional only when a number of conditions are satisfied.(16)

These conditions for anticompetitive distribution arrangements can be derived by considering an analogy to a conspiracy among distributors organized by a dominant manufacturer. In other words, the distributors and manufacturers work together to monopolize distribution, or both distribution and manufacturing, and share the industry monopoly profits.(17) An examination of the incentives facing an individual distributor to remain outside the cartel provides some insight as to the obstacles preventing monopolization through shelf space contracts.

For example, the distribution cartel analysis reveals that such monopolization requires economies of scale at the manufacturing level. If economies of scale in manufacturing are not significant, an individual distributor then has more to gain by remaining outside the conspiracy and contracting with a rival manufacturer. A competitive manufacturer in this case can survive at a limited scale and therefore does not require a large number of distributors. The strategy of depriving manufacturing rivals from access to distribution, and therefore creating or maintaining market power, is likely to be very costly. Like the incentive to cheat on a collusive agreement by expanding sales, a distributor would find it more profitable to contract with a rival manufacturer.

Conversely, where there are significant economies of scale in manufacturing, the potential exists for a stable, exclusionary distribution arrangement. However, a dominant manufacturer must cover enough distribution for a sufficient period of time so that a rival manufacturer is rendered unable to reach minimum efficient scale. A rival forced to reduce scale as a result of the scarcity of distribution operates at higher average costs and is therefore less able to discipline a price increase by the dominant manufacturer.

Economies of scale in distribution also play an important role in the economic analysis. Where there are significant economies of scale in distribution, a single distributor can supply a manufacturer of minimum efficient scale and, therefore, a dominant manufacturer would be forced to exclude a rival from each distributor in order to ensure that the rival was effectively excluded. In other words, any rival would defeat this exclusionary tactic by winning a single distribution contract. Even where contracts become available intermittently as a result of staggered expiration dates, it is unlikely that a manufacturer could exclude a competitive manufacturer from the market.

While many slotting contracts do not include exclusivity requirements and most involve relatively short-term retailer shelf space commitments,(18) exclusive and limited exclusive shelf space contracts frequently exist in circumstances where the required conditions for an anticompetitive effect are absent. For instance, slotting contracts are frequently used by manufacturers with small market shares, cover relatively small shares of total retail distribution, and involve grocery products that rarely involve substantial scale economies.(19) It is highly unlikely that these short term shelf space arrangements, even if they foreclosed a significant share of retail distribution from rivals, could successfully exclude rivals since competing manufacturers could openly compete for distribution and sign agreements with retailers.(20)

For example, since the limited exclusive arrangements in Gruma were relatively short-term, it seems highly unlikely that the contracts had the effect of anticompetitively foreclosing rivals.(21)

However, when a manufacturer with a large market share uses an exclusive shelf space contract it creates suspicions that the motivation and effect may be anticompetitive, in spite of the short-term nature of the contract, because an obvious pro-competitive rationale does not appear to exist for such contracts.(22)

When manufacturers and retailers negotiate restrictions on rival manufacturer shelf space placement as part of their slotting arrangements, as they did in Gruma, it is reasonable to presume that another brand would be stocked by the retailer without the agreed upon restriction. Such contracts, therefore, may be thought by some not to involve "competition on the merits," in the sense that the contracts do not involve solely the sale of particular shelf space to the highest bidder without concerns about shelf space supplied to rivals. However, in these circumstances where anticompetitive foreclosure is unlikely, if the competitive bidding process for shelf space leads to a full or partial exclusive restriction on rival manufacturer access to shelf space, it must be because retailers believe they can receive more for their shelf space than the lower consumer value from reduced variety. We now consider the case of category management contracts as another form of limited exclusive in the common case where some consumers' demand for a particular brand is high, and the retailer delegates to the manufacturer the power to determine the quantity of other, highly demanded products to be stocked.

  1. Category Management Contracts

Category management contracts are an alternative to the limited exclusive arrangements discussed in II.B. While category management refers more generally to the practice of organizing and managing retail distribution with the product category as the unit of analysis, these contracts also involve the retailer designating a particular manufacturer as the "category captain" in a product category, where the retailer grants exclusive rights to control offerings in the product category to the designated captain. These shelf space contracts leave it up to the category captain manufacturer to determine the quantities of other, highly demanded products that will be stocked, with the understanding that the chosen products will primarily, but not exclusively, be the category captain's products.(23) Category management contracts have become increasingly pervasive in retail distribution,(24) and antitrust challenges were mounted with respect to these contracts in both Gruma and Conwood.

Category management relationships exhibit significant variance across retailers and product categories.(25) The responsibilities delegated to the captain might include supplying category sales analysis, making shelf placement suggestions through "plan-o-grams," and making recommendations on product additions or deletions, promotions, and retail pricing subject to the retailer's right to terminate the relationship at will.

Competitive analysis of category management contracts has largely focused on the possibility that the manufacturer has superior information, and therefore, the delegation to the "captain" represents a vertical information-sharing arrangement.(26) However, informational asymmetries are an unlikely explanation for the pervasive adoption of "category captain" arrangements both because it is unlikely that manufacturers have superior information about the demand for rival products and because if each manufacturer has superior information regarding distribution of its own product, one would expect the retailer to collaborate with multiple manufacturers rather than delegating these responsibilities to a single manufacturer.

While the FTC Report concedes that "category management can produce significant efficiencies that will benefit retailers, manufacturers and consumers,"(27) it also expresses two primary competitive concerns. The first is that the category manager relationship might facilitate collusion between either manufacturers or retailers.(28) The second concern is that the captain may use its position to effectively exclude or significantly disadvantage competitors, exposing consumers to the risk of decreased product variety or increased prices.(29) Antitrust challenges to category management contracts, such as those in Conwood and Gruma, have been restricted to the latter anticompetitive theory.

It is unlikely that the anticompetitive exclusion of rivals explains the widespread adoption of category management contracts. The conditions under which restrictive distribution contracts may produce anticompetitive outcomes, discussed above, are not likely to be satisfied by most category management contracts. Category management contracts are generally of short duration and terminable at will. Further, category captains frequently do not possess antitrust market power. For instance, several competing manufacturers without significant shares in the moist snuff market at issue in Conwood successfully were designated as category captain by major retailer chains. Anticompetitive theories of category management contracts also do not address the fundamental economic question of why a retailer delegates the shelf space allocation decision rights to a manufacturer since exclusion of rival manufacturers requires long-term control of shelf space, which can be achieved without the delegation of shelf space allocations rights to a category manager.

We now turn to our procompetitive explanation of category management contracts as an alternative solution to the promotional shelf space contracting problem that arises when some customers have high demand for particular rival brands. The implicit contractual understanding regarding performance is not as clear under these conditions as in the more standard slotting shelf space contract where retailers supply a particular quantity, location, or exclusive (or particular fraction of) shelf space. Because performance cannot be precisely defined, the possibility of non-performance on the margin exists on both sides of the transaction.(30) The efficient shelf space contract under these conditions generally involves some form of limited exclusive.

Category management contracts, like limited exclusives, therefore offer retailers the ability to commit its marginal consumers to manufacturers and increase the value of its shelf space when consumers value specific rival brands. The manufacturer, therefore, is assured of receiving the limited exclusive it has paid for and the retailer relies on the manufacturer brand name to assure manufacturer performance, that is, to assure that overall retailer profit is not reduced by the category captain's decisions that make rival offerings too restrictive.(31)

A separate question exists as to why this particular form of limited exclusive involves shifting shelf space decisions, and therefore the opportunity to violate the implicit contract, to the manufacturer, while the retailer becomes the policer of the contract. This shift may be efficient when the manufacturer has a greater brand name than the retailer. Furthermore, it is likely that the retailer has lower policing costs than the manufacturer. While it may be costly for the manufacturer to monitor and detect subtle violations of the implicit promotional agreement by the retailer in the form of rearranging rival products or failure to promptly restock sold out facings, the retailer may more efficiently detect manufacturer violations such as restricting rival offerings in a manner that reduces overall retailer profit.

Category management contracts, like most shelf space contracts, are almost always self-enforced.(32) Transactors will continue to perform in a manner consistent with the implicit understanding so long as the expected premium stream earned by the party over the life of the agreement is greater than the gains from cheating. The retailer's gains from short-run cheating on the category management relationship, Wr1, are defined by the gains from the retailer from promoting a rival product or supplying less than the agreed upon level of shelf space to the retailer. Designating a category manager can be thought of as reducing the retailer's incentive to cheat by placing the shelf-space decisions in the hands of the manufacturer, thereby increasing the probability that nonperformance will be detected and increasing the probability that the relationship remains within the self-enforcing range over the maximum number of possible contingencies. On the other hand, Wr2, the premium stream earned by the retailer, consists of the per unit time or per unit sale payments paid by the manufacturer over the duration of the agreement. From the retailer's perspective, designating a category manager allows the retailer to earn greater manufacturer payments than would prevail without the designation while minimizing the costs associated with the reduction in product variety. The retailer will continue to perform as long as Wr2 > Wr1.

Manufacturer cheating may be defined as the supply of less than the "agreed upon" level of product variety by restricting rivals. The gains from the manager associated with short-run cheating, Wm1, are defined by the extra profits earned by the manager from shifting sales from restricted rivals to its own product. The magnitude of the sanction that can be imposed on the manufacturer, Wm2, is the "extra" profits earned by the manufacturer over the course of the relationship from increased promotional sales. Because the manager stands to lose these profits upon termination, the manager will perform so long as Wm2 > Wm1.

The mere fact that an agreement is designed to be self-enforcing, of course, does not guarantee that the category manager will always perform in accordance with its obligations. Transactors attempt to design contract terms to minimize the probability of future non-performance, but unanticipated events can take performance outside the self-enforcing range in some circumstances. This appears to be what was alleged to have occurred in Conwood where United State's Tobacco's ("UST") abuses of its category manager position, including unauthorized removal and destruction of competitors' display racks and product, were found in violation of the Sherman Act.(33) Section IV analyzes the antitrust law of category management, and in particular, the legal duties faced by category managers in the wake of Conwood.

  1. Antitrust Analysis of Category Management

Antitrust scrutiny of shelf space arrangements has been largely limited to full or partial exclusive contracts in which it is alleged that the manufacturer is paying too much for shelf space over a significant period of time, or unnecessarily restricting rivals' access to shelf space. The primary contribution of our economic analysis to the antitrust treatment of category management contracts is that these arrangements are properly thought of as limited exclusives. Because a category management shelf space contract is a limited rather than a full exclusive, it is an inherently less restrictive agreement in terms of its potential to foreclose rivals. Antitrust law, however, paradoxically appears to impose greater legal duties on category managers than manufacturers purchasing exclusive shelf space.

Exclusive dealing jurisprudence has generally adopted a standard largely consistent with identifying those arrangements which satisfy the conditions for anticompetitive outcomes discussed in Section III. This analysis requires a demonstration of the defendant's market power, quantitative and qualitative foreclosure sufficient to generate an anticompetitive effect in light of the duration of the contracts and entry conditions, as well as consideration of plausible procompetitive justifications.(34)

Conwood adopts a relatively unsophisticated economic approach which overstates the competitive threat of category management arrangements relative to fully exclusive shelf space arrangements.(35) While the conditions necessary for a category management contract to result in harm to competition are analytically identical to those for exclusive shelf space arrangements, Conwood fails to provide an adequate framework for identifying which category management relationships merit antitrust scrutiny.

Confusion regarding the economic function of category management has resulted in debate on the most basic levels of antitrust analysis. For example, commentators have recently questioned whether category management should be characterized as a horizontal or vertical restraint.(36) Proponents of horizontal analysis assert that post-Sylvania antitrust jurisprudence protects a manufacturer's interest in managing the retailing of its own product, but not those of its rivals.(37) Under this view, category captain recommendations regarding rivals' products are not protected and should be treated as horizontal for antitrust purposes. As a legal matter, Glazer, Henry, and Jacobson persuasively respond that the Supreme Court unequivocally adopted the formal horizontal/vertical distinction in Sharp, and reject the view that any effect on interbrand competition is sufficient to justify horizontal treatment.(38)

Our economic analysis of category management contracts as limited exclusive shelf space arrangements supports the view that these contracts are properly viewed as vertical restraints for antitrust purposes. While both full and limited exclusive arrangements allow retailers to maximize shelf space value by increasing the elasticity of demand faced by manufacturers bidding for shelf space, category management contracts arise only when consumer demand for product variety within a category renders a full exclusive too costly. From an economic standpoint, limited exclusives involve less influence over rival product offerings than full exclusives which are analyzed subject to a rule of reason analysis. Because category management is a limited exclusive which is inherently less of a competitive threat than exclusive dealing, these agreements and communications associated with them should also be analyzed as vertical arrangements in the absence of evidence of a conspiracy between competitors. As discussed below, however, even greater antitrust scrutiny has been applied to the conduct of category managers than monopolists utilizing full exclusives.

    1. Category Management in the Courts: Conwood Co. v. United States Tobacco Co.

Conwood involved UST, the leading manufacturer in the $1.7 billion U.S. moist snuff market, and its product distribution campaign. UST held a 77% share of the moist snuff market while Conwood, the plaintiff, held approximately 13%.(39) UST manufactures the "Skoal" and "Copenhagen" brands, while Conwood manufactures the "Kodiak" and "Cougar" brands.(40) UST had historically dominated the market for most of the 1970s and 1980s, only to experience a decrease in market share during the 1990s in the wake of annual price increases from 1979-88 and increased entry.(41)

Product placement and distribution is a key dimension of competition in the moist snuff market. In particular, in-store and "point of sale" advertising and promotional effort in the moist snuff market became increasingly important after the Master Tobacco Settlement imposed new restrictions on tobacco advertising.(42) During the 1990s, UST launched a new distribution strategy which included the aggressive pursuit of "exclusive" racks. An "exclusive" rack does not refer to a rack containing only the product of a single supplier. Rather, it refers to a rack that exclusively advertises a single moist snuff supplier, but includes rivals' products.(43) A key competitive benefit of obtaining an exclusive rack is that the supplier may make product placement decisions on the rack subject to retailer approval. That is, a supplier with an exclusive rack is responsible for many of the functions assigned to a category manager.

Conwood's monopolization claim not only challenged UST's strategy of seeking exclusive racks, but also alleged abuse of its category manager position.(44) Specifically, the most troubling of these allegations were that UST misled retailers, removed Conwood display racks without retailer approval, and even destroyed rivals' racks and product.(45) At trial, Conwood produced some evidence supporting its allegations that UST salespersons had indeed removed Conwood display racks without authorization.(46) Some UST employees testified that they had orders from supervisors to remove rival racks and replace them with UST racks.(47) On this and other evidence, the jury returned the $350 million verdict (before trebling) in Conwood's favor after just four hours of deliberation.(48)

UST raised three major objections to the district court's decision on appeal to the Sixth Circuit: (1) the product destruction and removal amounted to "no more than isolated sporadic torts"(49) rather than a widespread and systematic strategy sufficient to produce an antitrust violation; (2) Conwood had not established a causal link between any of UST's business practices and antitrust injury;(50) and (3) Conwood failed to show sufficient foreclosure from the market for shelf space. The Sixth Circuit rejected the first two arguments out of hand, finding that UST's conduct qualified as one of the rare, "gross cases"(51) of tortious business practices sufficient to trigger antitrust liability.

The Sixth Circuit's foreclosure analysis made clear that the court would not handle the claims in a manner consistent with the line of cases challenging exclusive dealing arrangements under Sections 1 and 2 of the Sherman Act or Section 3 of the Clayton Act. The Sixth Circuit rejected UST's argument that a particularized showing of foreclosure separating out instances of authorized removal from unauthorized removal was necessary because Conwood's claims involved product destruction and the abuse of a category captain relationship rather than simple exclusive dealing, and because Conwood's claims were not adjudicated under Clayton Act § 3.(52)

Conwood's analysis commits two major analytical flaws. The first is that Conwood penalizes UST for its role as category manager without clearly delineating the duties imposed on similarly situated dominant firms, and therefore paradoxically imposes a more exacting antitrust standard on dominant firms operating as category managers than on those adopting exclusives. More specifically, the court reasons that UST's abuses as category manager, including product destruction, excuse rigorous analysis of foreclosure and anticompetitive effects. The second flaw, discussed in IV.B, is that the court incorrectly concludes that the evidence presented at trial was sufficient to support an inference of anticompetitive effect.

Contrary to the general principle that a monopolist has no duty to assist rivals,(53) Conwood implicitly imposes a fiduciary-like obligation on category managers to assist rival brands' distribution efforts by dramatically reducing the plaintiffs' burden in challenging the manager's conduct. Specifically, Conwood does not require any showing of foreclosure from distribution in order to assess the likelihood of anticompetitive effects. The court explicitly rejects the notion that a showing of foreclosure was necessary because Conwood does not "merely challenge the exclusive agreements UST entered into" but also includes allegations of rack destruction, misrepresentation, and abuse of category management position to disadvantage rivals.(54) Further, the court also did not require Conwood to distinguish injuries flowing from UST's superior ability to obtain exclusive agreements or other forms of legitimate competition from those caused by conduct violating the antitrust laws because "UST and not retailers controlled facing decisions and that in making those decisions, UST sales representatives purposely attempted to bury Conwood's product in UST racks."(55)

One interpretation of the Conwood standard is that in the absence of a business justification for the conduct at issue, alleged abuses of category management relationships will be sufficient to support a violation of Section 2 without any evidence of substantial foreclosure or anticompetitive effect. Perhaps this standard will eventually be limited to instances in which such abuses are coupled with allegations of activity, such as product destruction, that clearly falls outside the scope of "competition on the merits." However, no such limitation currently exists. The problem with such an approach is that it carries an unacceptably large risk of condemning efficiency-enhancing conduct or "hybrid" conduct involving at least some allegations of tortious activity but which is incapable of producing an injury that the antitrust laws were designed to prevent.

We now turn to the question of whether, based upon the evidence presented at trial in Conwood, UST's conduct was sufficient to demonstrate an anticompetitive effect under the more demanding rule of reason standard conventionally applied in exclusive dealing cases under Section 1 of the Sherman Act and Section 3 of the Clayton Act, or even the less stringent Section 2 foreclosure requirement as applied Microsoft.

    1. There Was Insufficient Evidence to Conclude that UST's Conduct Was Anticompetitive

Section II discussed the conditions under which a dominant manufacturer may foreclose competing manufacturers and potential entrants by tying up a sufficient mass of a critical input through contractual arrangements such that the manufacturer or potential entrant is unable to achieve minimum efficient scale. It is obvious that at least some portion of UST's conduct was not "competition on the merits." Nonetheless, the question remains whether the conduct was capable of generating an anticompetitive effect. We argue that at least one necessary condition for anticompetitive harm, foreclosure sufficient to deprive rivals' ability to achieve minimum efficient scale, was likely absent in Conwood.

A dominant firm's distribution strategy is capable of harming competition only if rivals are foreclosed from access to shelf-space at a level sufficient to deprive them from achieving minimum efficient scale. Though this requirement has been largely incorporated into the antitrust analysis of exclusionary distribution contracts, both the district court and the Sixth Circuit eschewed any meaningful quantitative or qualitative foreclosure analysis. Instead, the court assumed that UST's aggregate was sufficiently widespread to create competitive injury.(56) While the record supports Conwood's allegations that UST engaged in intentionally tortious conduct, there is very little evidence of competitive foreclosure and the possibility of anticompetitive effect.

For example, UST successfully bid for exclusive racks for moist snuff at less than 10% of retail outlets.(57) These exclusive racks featured UST product and signage, but also included the products of rivals such as Conwood, Swisher, and Swedish Match. The low percentage of UST exclusive racks implies that any advantage bestowed upon UST as category captain was limited to a small fraction of retail outlets. Furthermore, Conwood's distribution rate measured by percentage of retail outlets offering its products was 81percent,(58) compared to UST's 87percent distribution rate.(59) Even under the unrealistic assumption that Conwood was foreclosed from each of these outlets as the result of UST's conduct, Conwood was only foreclosed from 19percent of retail outlets, a foreclosure rate prone to summary judgment.

The number of facings or "slots" may be a more appropriate unit of analysis because it addresses the possibility that although Conwood was nominally present in the store, UST's conduct anticompetitively restricted Conwood's shelf space. For example, UST might replace a Conwood rack with six Conwood facings with its own UST rack with only one or two Conwood facings. A complete foreclosure analysis may require a calculation of shelf space facings available to Conwood as a result of UST's conduct. The Sixth Circuit, however, did not require Conwood to produce evidence of shelf space or rack foreclosure at this level of detail.

The foreclosure analysis is further complicated by the allegations of unauthorized product destruction and removal mixed with rack removal authorized by the retailer. While the antitrust laws are clear that tortious conduct may violate Section 2, such conduct must be demonstrated to have an anticompetitive effect.(60) There are clearly were some instances of unauthorized rack removal,(61) but a substantial fraction of the total rack removal appears to have occurred with the retailer's permission.(62) There is no doubt that Conwood made significant investments to replace its removed racks, whether the removal was authorized or otherwise. For example, one Conwood executive testified that the cost of replacing racks amounted to as much as $100,000 per month,(63) and that Conwood employees spent about 50% of their time replacing removed racks, amounting to approximately 20,000 racks per month.(64) Such repeated misconduct over time, however, would have likely resulted in the termination of UST as category manager in retail outlets because such non-performance would be readily observable. However, there is no evidence of widespread retailer termination of UST as category manager over the relevant time period. To the contrary, a large organization of chain stores bestowed UST with an award for excellence in category management during the same time period based on a survey of retailers.(65) However, even attributing all of these Conwood costs to the replacement of unauthorized rack removal, the competitive significance of these costs are de minimis.(66)

It is unclear how much of these costs are attributable to unauthorized removal of racks.(67) One possible explanation consistent with this evidence is that repeated and systematic unauthorized rack removal at retail outlets went on unchecked and without terminating UST as category manager because retailers did not have another choice as category manager. The record contradicts this explanation. Moist snuff rivals such as Swisher and Swedish Match, as well as distributors, served as moist snuff category manager on major accounts.(68) A second and more plausible explanation of the evidence is that the UST unauthorized removal involved sporadic abuses by UST employees and not a systematic and widespread failure of the contractual relationship between UST and retailers.(69) When retailers testified that UST employees removed product without authorization, the result was frequently that store managers would get a replacement rack from Conwood or request that UST remove the new rack.(70)

Despite the state of evidence insufficient to support an inference of anticompetitive effect, the Sixth Circuit affirmed a $1.05 billion award, the largest verdict in the history of antitrust law. The estimates proffered by Conwood's expert witness and affirmed by the Sixth Circuit have been heavily criticized for their failure to provide any meaningful measure of antitrust damages.(71) The lynchpin of Conwood's damages analysis was a regression analysis of the relationship between Conwood's market share in 1990 and its market share growth from 1990-97 across states. The expert then assumed that Conwood's lower growth in states with low market shares in 1990 was caused solely by UST's conduct, based upon the hypothesis, devoid of economic logic or theoretical underpinnings, that UST's illegal conduct would differentially impact its subsequent share growth in states where it started with a sufficiently "low" market share, designated on an ad hoc basis to be 20 percent.(72) Indeed, the Sixth Circuit credits this testimony as demonstrating that the expert "found a statistically significant difference in Conwood's market share between those states in which Conwood has a foothold and those in which it did not."(73) It was this difference that the Sixth Circuit relied upon to conclude that Conwood would have grown at some constant rate in all of the "low share" states but for UST's illegal conduct.

Conwood's damages analysis was inherently incapable of measuring any potential antitrust injury associated with UST's illegal conduct, and commentators have uniformly agreed that the testimony should have been precluded by Daubert.(74) The damages analysis fails to distinguish the effect on competition caused by UST's illegal conduct from any effects associated with other causes, including legitimate forms of competition which would not violate Section 2 or other factors that might explain the variation in growth rates across states. In fact, Conwood's analysis does not control for any differences in UST's marketing and distribution practices across states.

In addition to legal conduct, another important intervening cause of Conwood's market share losses appears to be its own failure to compete for product distribution. Conwood failed to participate in the plan-o-gram competition offered by Wal-Mart.(75) Conwood also complained that a UST discount program granting a .3% discount to retailers providing UST products preferred shelf space was able to secure commitments from 37,000 retailers,(76) despite the fact that Conwood failed to offer similar discounts to retailers.(77) Furthermore, Conwood executives testified that they considered and ultimately rejected an investment into category management capabilities, unlike each of its major moist snuff competitors.(78)

In sum, there is very little evidence to support the inference, affirmed by the Sixth Circuit, that UST's distribution practices harmed competition or was likely to generate an anticompetitive effect. The available evidence does not suggest foreclosure levels sufficient to trigger liability under modern antitrust standards applied to exclusionary distribution arrangements under Sections 1 or 2 of the Sherman Act or Section 3 of the Clayton Act. Nor does the evidence support the heavily criticized damages estimates affirmed by the Sixth Circuit.

The Sixth Circuit's failure to require Conwood to demonstrate substantial foreclosure, the "hybrid" nature of Conwood's claims mixing tortious behavior with legal conduct, and the lack of an obvious procompetitive justification for UST's conduct as category manager and rack destruction, combined to produce the perverse result which imposes more stringent antitrust duties on category managers than monopolists distributing their products through more restrictive exclusive requirements. Our analysis presents a procompetitive justification for category management arrangements likely to be challenged in Conwood's wake, and suggests that the potentially exclusionary effects from abuses of these contracts should be analyzed in the same manner as limited exclusives.

  1. Conclusion

This paper analyzes the economic forces at work in exclusive retail shelf space contracts. Specifically, we focus on category management contracts as a form of limited exclusive which shifts responsibility for shelf space decisions to the manufacturer and allocates the policing function to the retailer. When consumers are brand loyal and not likely to switch to competing brands in response to a reduction in variety within a product category, exclusivity is a natural consequence of competition for shelf space and allows a retailer to obtain greater value for its shelf space by effectively promising to transfer its customers' purchases to the selected manufacturer. In this way, the retailer's commitment to exclusivity increases the elasticity of demand facing manufacturers bidding for the retailer's shelf space, generating greater payments for shelf space which are ultimately passed on to consumers.

In the common case when consumer demand for variety within the product category or for a particular brand is high, this analysis explains why the efficient shelf space contract involves a limited exclusive. Category management contracts are best viewed as a form of a limited exclusive, where the manufacturer determines the quantities of other, highly demanded products to be stocked. Category management is likely to be adopted when retailer performance cannot be specified precisely and, therefore, there is a possibility of hold-ups on both sides of the transaction. Limited exclusives, including category management, offer the retailer an alternative method of committing promotional sales to a single manufacturer, thereby intensifying competition for shelf space by increasing the elasticity of demand faced by each manufacturer and increasing the retailer's return on its shelf space.

The current antitrust scrutiny of full and limited exclusive shelf space contracts and category management relationships threatens to deter the use of these contracts resulting from competition for shelf space, which benefit consumers as a result of inter-retailer competition. Our economic analysis suggests that allegations of exclusionary abuses of category management relationships should be analyzed similarly to limited exclusives. Furthermore, our pro-competitive explanation of category management contracts suggests that modern category management jurisprudence, as represented by Conwood, perversely imposes greater antitrust duties on category managers than on those firms implementing more restrictive distribution arrangements.


Areni, Charles, Dale Duhan & Pamela Kiecker [1999]. "Point-of Purchase Displays, Product Organization, and Brand Purchase Likelihoods," 27(4) Journal of the Academy of Marketing Science 428-41.

Basuroy, Suman, Mantrala, Murali, and Walters, Rockney [2001], "Impact of Category Management on Retail Prices and Performance: Theory and Evidence." 65 Journal of Marketing 16-32.

Desrocher, Debra M., Gundlach, Gregory T., and Foer, Albert A. [2003], "Analysis of Antitrust Challenges to Category Captain Arrangements," 22 Journal of Public Policy and Marketing 201.

Dreze, Xavier, Stephen P. Hoch & Mary E. Purk [1994]. "Shelf Management and Space Elasticity," 70 Journal of Retailing 301.

Federal Trade Commission Staff Report [2001], "Report on the Commission Workshop on Slotting Allowances and Other Marketing Practices in the Grocery Industry."

Kenneth Glazer, Henry, Brian R. Henry, and Jacobson, Jonathan Jacobson [2004], "Antitrust Implications of Category Management: Resolving the Horizontal/ Vertical Characterization Debate," The Antitrust Source (July 2004).

Granitz, Elizabeth & Klein, Benjamin [1996], "Monopolization by Raising Rivals' Costs": The Standard Oil Case, 39 Journal of Law & Economics 1.

Jacobson, Jonathan [2002], "Exclusive Dealing, 'Foreclosure,' and Consumer Harm," 70 Antitrust L.J. 311.

Kaye, David [2001], "The Dynamics of Daubert Methodology, Conclusions, and Statistical Fit in Econometric Studies" 87 U. Va. L. Rev. 1933.

Kaye, D.H. [2003], "Adversarial Econometrics in United States Tobacco v. Conwood Co.," 43 Jurimetrics 343.

Klein, Benjamin [2003], "Exclusive Dealing as Competition for Distribution 'On the Merits,'" 12 George Mason University Law Review 119 (2004).

Klein, Benjamin & Leffler, Keith [1981], "The Role of Market Forces in Assuring Contractual Performance," 89 Journal of Political Economy 615.

Klein, Benjamin & Lerner, Andres [2006], "How Exclusive Dealing Prevents Free-Riding and Creates Dealer Loyalty," unpublished paper.

Klein, Benjamin & Murphy, Kevin M. [1988], "Vertical Restraints as Contract Enforcement Mechanisms," 31 Journal of Law & Economics 265-97.

Klein, Benjamin & Kevin M. Murphy [2006], "Exclusive Dealing Intensifies Competitive Bidding for Distribution," unpublished paper.

Klein, Benjamin & Wright, Joshua, [2007], "The Economics of Slotting Contracts," Journal of Law and Economics, forthcoming.

Leary, Thomas [2004], "A Second Look at Category Management," available at http://

Niraj, Rakesh and Narasimhan, Chakravarthi [2006], "Vertical Information Sharing in Distribution Channels," (unpublished working paper, 2006)

Rennhoff, Adam [2004], "Paying for Shelf Space: An Investigation of Merchandising Allowances in the Grocery Industry (July 2004).

Steiner, Robert [2001], "Category Management--A Pervasive New Horizontal/ Vertical Format," 15 Antitrust 77.

Wright, Joshua D., [2006], "Antitrust Law and Competition for Distrbution," 23 Yale Journal on Regulation 169 (2006).



_ Klein: Professor Emeritus of Economics, UCLA and Director, LECG, LLC; Wright: Assistant Professor, George Mason University Law School. We appreciate research assistance by Brandy Wagstaff.

1. See, e.g., El Aquila Food Prods. v. Gruma Corp., 301 F. Supp. 2d 612 (S.D. Tex. 2003), aff'd, 131 F. App'x 450 (5th Cir. 2005); R.J. Reynolds Tobacco Co. v. Philip Morris, Inc., 199 F. Supp. 2d 363 (M.D.N.C. 2002), aff'd per curiam, 67 F. App'x 810 (4th Cir. 2003); Conwood Co. v. United States Tobacco Co., 290 F. 3d 768 (6th Cir. 2002); American Booksellers Ass'n, Inc. v. Barnes & Noble, Inc., 135 F. Supp. 2d 1031 (N.D. Cal. 2001); Intimate Bookshop, Inc. v. Barnes & Noble, Inc., 88 F. Supp. 2d 133 (S.D.N.Y. 2000); FTC v. McCormick, FTC Dkt. No. C-3939 (2000).

2. See Federal Trade Comm'n, Report on the Federal Trade Commission Workshop on Slotting Allowances and Other Marketing Practices in the Grocery Industry 46-55 (2001) [hereinafter FTC Report] (discussing category management).

3. These restrictive shelf space contracts can be distinguished from single product exclusive dealing observed, for example, in automobile distribution. See Benjamin Klein & Andres Lerner, How Exclusive Dealing Prevents Free-Riding and Creates Dealer Loyalty (unpublished paper, 2006).

4. Benjamin Klein & Joshua D. Wright, The Economics of Slotting Contracts, J.L. & Econ. (forthcoming 2007).

5. Benjamin Klein & Kevin M. Murphy, Exclusive Dealing Intensifies Competitive Bidding for Distribution (unpublished paper, 2006).

6. El Aquila Food Prods v. Gruma Corp., 301 F. Supp. 2d at 615. Similar allegations appear in a complaint against Ocean Spray Cranberries which was later settled. See Complaint at ¶¶ 9, 54, Northland Cranberries, Inc. v. Ocean Spray Cranberries, Inc. (D.D.C. Nov. 8, 2002) (alleging that Ocean Spray abused its position as category captain for shelf-stable juices).

7. El Aquila Food Prods. v. Gruma Corp., 301 F. Supp. 2d at 621.

8. The analysis in section II.A appears in Klein and Wright, supra note 4. The original statement of the problem of insufficient retailer supply of promotion appears in Benjamin Klein and Kevin M. Murphy, Vertical Restraints as Contract Enforcement Mechanisms, 31 J.L. & Econ. 265 (1988).

9. Many marketing studies have confirmed that shelf-space positioning increases the sales of the featured product. Adam Rennhoff, Paying for Shelf Space: An Investigation of Merchandising Allowances in the Grocery Industry (working paper, on file with author, July 2004); Xavier Drezè, et al., Shelf Management and Space Elasticity, 70 J. Retailing 301 (1994); Charles Areni, et al., Point-of-Purchase Displays, Product Organization, and Brand Purchase Likelihoods, 27 J. Acad. Mktg. Sci. 428 (1999).

10. It is generally reported that the magnitude and frequency of slotting allowances, or per unit time payments from manufacturers to retailers for the provision of shelf space, has increased in recent years. FTC Report, supra note 2, at 11. Klein and Wright show that the growth in slotting is consistent with increased product differentiation and mark-ups in products sold in grocery retail outlets as measured by manufacturer "value-added," defined as the difference between costs of production and the value of shipments. Klein & Wright, supra note 4.

11. Compensation for promotional shelf space is therefore a substitute for other vertical restraints also used by manufacturers to pay retailers for supplying the desired level of promotional effort, such as exclusive territories or resale price maintenance. See Klein & Murphy, supra note8.

12. One concern raised regarding per unit time payments such as slotting allowances is that they will not be passed on to consumers. This analysis incorrectly assumes that retail competition is limited to a single margin, the price of the category manager's good. Given free entry into grocery retail, any payments above the competitive rate of return can be expected to be competed away on some dimension, most likely price discounts on products likely to increase grocery traffic, or other in-store amenities likely to induce inter-retailer demand effects. The evidence suggests that retailer profit margins have not increased over the time period during which category management and shelf space payments surged in popularity, suggesting that these payments are passed on to consumers as the result of retail competition. See Klein & Wright, supra note4.

13. See cases cited supra note 1.

14. Klein & Murphy, supra note 5.

15. FTC Report, supra note 2, at 34-41.

16. The following section relies on the analysis presented in Benjamin Klein, Exclusive Dealing as Competition for Distribution 'On the Merits,' 12 Geo. Mason L. Rev. 119 (2004).

17. This analysis has been shown to describe the role of Standard Oil in enforcing the collusively agreed upon individual railroad market shares. See Elizabeth Granitz & Benjamin Klein, Monopolization by "Raising Rivals' Costs": The Standard Oil Case, 39 J.L. & Econ. 1 (1996).

18. Interviews with manufacturers and retailers indicate that the most common time period for a new product stocking commitment is a minimum of six months. FTC Study, supra note 2, at iii n.14; see also id. at 11 (citing Sussman, Tr. at 83-84). Slotting contracts that deal with stocking commitments for established products are usually one year in duration. Id. at 57. Other shelving commitments, for example, the display of a particular product at the end of an aisle, may be substantially shorter-term, sometimes covering only a week. See the description of Frito-Lay's marketing practices in Department of Justice Memorandum (May 16, 1996), available at

19. See Klein & Wright, supra note4, for a more complete discussion.

20. Several courts have established a safe harbor for short-term exclusive agreements and those that are terminable on short notice. See, e.g., Roland Mach. Co. v. Dresser Indus., 749 F.2d 380, 395 (7th Cir. 1984) (exclusive dealing contracts terminable in less than one year are presumptively lawful under Section3 of the Clayton Act ); Omega Envtl., Inc. v. Gilbarco, Inc., 127 F.3d 1157, 1162 (9th Cir. 1997) (citing Roland Machinery and stating that the "short duration and easy terminability of these agreements negates substantially their potential to foreclose competition"), cert. denied, 525 U.S. 812 (1998); and R.J. Reynolds Tobacco Co. v. Philip Morris, Inc., 199 F. Supp. 2d 363, 391 (M.D.N.C. 2002), aff'd per curiam, 67 F. App'x 810 (4th Cir. 2003) (concluding that because Philip Morris agreements with retailers were terminable at will with thirty days notice, "retail product and display space are subject to uninterrupted competitive bidding, and Plaintiffs are not substantially foreclosed from the relevant market").

21. The longest agreement referred to by plaintiffs in Gruma was a two-year commitment granting Gruma exclusivity in the branded tortilla space in a single chain. See Appellants' Brief, El Aquila Food Prods. v. Gruma Corp., 301 F. Supp. 2d 612 (S.D. Tex. 2003), aff'd, 131 F. App'x 450 (5th Cir. 2005) (No. 04-20125, June 16, 2004) (available from authors). Moreover, significant economies of scale do not exist with regard to the manufacture of tortillas. Many small tortilla manufacturers continued to exist in the face of Gruma's contracts and some continued to sell their products on shelves next to Gruma's brands. 301 F. Supp. 2d at 630. In addition, Gruma continued to compete against retailer private label in-store tortilla machines that had relatively low production rates. Id. at 624.

22. The FTC Report, supra note 2, at 39, recognizes the prevention of "classic dealer free-riding," where rival retailers take advantage of retail promotional services, as a possible efficiency justification for exclusive arrangements. The FTC Report, id. at 45, also mentions the possibility that an exclusive shelf space arrangement "commits the retailer to a single brand of product, and so may induce the retailer to devote greater attention to that brand."

23. Five of the seven retailers surveyed in a Federal Trade Commission Staff Study reported that they used category management contracts for some products. Federal Trade Comm'n Staff Study, Slotting Allowances in the Retail Grocery Industry: Selected Case Studies in Five Product Categories (Nov. 2003). Category management has been employed in several retail trades for years, but is relatively new to the grocery retail industry. FTC Report, supra note 2, at 47.

24. A recent study showed that 78% of department stores, 74% of discount stores, and 45% of supermarkets deploy category management. Chain Store Age (March 2000). A more recent estimate suggests that category management has been adopted in 75 percent of supermarkets and over 50 percent of drug and department stores. Chain Store Age/ KPMG Ninth Annual Survey of Inventory Management (Dec. 2001).

25. FTC Report, supra note 2, at 48.

26. The FTC Report notes that "the manufacturer may know things like the times of year when a product will sell best, the kind of promotions that are most effective in moving the product, or the kinds of complementary goods that might be advantageously displayed in adjacent space." Id. at 48. This argument has been echoed in the marketing literature. See, e.g., Debra Desrocher, et al., Analysis of Antitrust Challenges to Category Captain Arrangements, 22 J. Pub. Pol'y & Mktg. 201, 207 (2003); Rakesh Niraj & Chakravarthi Narasimhan, Vertical Information Sharing in Distribution Channels (unpublished working paper, 2006).

27. FTC Report, supra note 2, at 54.

28. Id. at 51-52. See also Desrocher, et al., supra note 26; Robert Steiner, Category Management--A Pervasive New Horizontal/ Vertical Format, 15 Antitrust 77 (Spring 2001).

29. FTC Report, supra note 2, at 51 ("a captain that is able to control decisions about product placement and promotions could hinder the entry or expansion of other manufacturers, leading to less variety and possibly higher prices"). One study found that the introduction of category management, rather than brand-level management, in the laundry detergent segment at a single retailer was associated with higher average retail prices, reduced wholesale prices, lower sales, and higher retailer category profits. See Suman Basuroy, et al., Impact of Category Management on Retail Prices and Performance: Theory and Evidence, 65 J. Mktg. 16 (2001).

30. In particular, the retailer has the ability to supply greater shelf space to competing brands than implicitly contracted for while the manufacturer has the ability to restrict the supply of competing brands more than implicitly contracted for. While the parties may be able to contractually specify the quantity and quality of promotional shelf space, retailers retain the ability to cheat on the implicit agreement by "selling the same space twice," for instance, by shifting some promotional sales of the featured brand to rivals with subtle changes in the allocation of rival products or failing to promptly re-stock sold out facings of the featured product.

31. One way to think about a category manager designation is as the retailer providing an "exclusive promotion" contract rather than an exclusive on the retailer's shelf space. The implicit understanding between the manufacturer and retailer is essentially that the retailer will only promote the manager's product through the provision of a greater quantity and quality of shelf space. In turn, the manager's obligation is to pay for the shelf space with a per unit time or per unit sale contract, and to make recommendations to the retailer or design shelf space allocations that will not harm the retailer in terms of reduced variety. The designation of a category manager or use of a contractual device to ensure the satisfaction of those consumers with a specific demand for another product are alternative methods for limiting the costs associated with reduced product variety while obtaining the benefits of exclusive promotion, specifically, increasing the value of the promotional effort sold to the manufacturer.

32. See Benjamin Klein & Keith Leffler, The Role of Market Forces in Assuring Contractual Performance, 89 J. Pol. Econ. 615 (1981).

33. UST non-performance was unlikely because retailers could easily detect UST non-performance, such as failing to display products in the manner desired or removing racks from the store, and terminate the agreement. Likewise, there is little evidence that retailers were systematically not performing with regard to their shelf space obligations to UST. Rather, the record suggests that the majority of category management relationships between UST and retailers did not involve unauthorized product destruction or rack removal. Infra Section IV.B.

34. See, e.g., R.J. Reynolds Tobacco Co. v. Philip Morris Inc., 199 F. Supp. 2d 362 (M.D.N.C. 2002), aff'd per curiam, 67 F. App'x 810 (4th Cir. 2003); Omega Envtl., Inc. v. Gilbarco Co., 127 F.3d 1157, 1163-64 (9th Cir. 1997). The recently released FTC Report proposes a similar analytical framework for exclusionary distribution cases, focusing on whether the defendant's conduct deprived a rival from achieving minimum efficient scale and harmed competition. FTC Report, supra note 2, at 36-40. See generally Joshua D. Wright, Antitrust Law and Competition for Distribution, 23 Yale J. Reg. 169 (2006).

35. See supra note 34; Herbert Hovenkamp, The Antitrust Enterprise: Principle and Execution 180 (Cambridge Press 2005) (describing Conwood as "deeply troublesome and offensive to antitrust policy").

36. See, e.g., Thomas Leary, A Second Look at Category Management 2-3 (May 17, 2004), available at http://www. (arguing that category management is a horizontal agreement); cf. Kenneth Glazer, Brian R. Henry & Jonathan Jacobson, Antitrust Implications of Category Management: Resolving the Horizontal/ Vertical Characterization Debate, The Antitrust Source (July 2004) (category management is appropriately characterized as a vertical restraint).

37. See Leary, supra note 36.

38. See Glazer, Henry & Jacobson, supra note 36 (quoting Business Electronics Corp. v. Sharp Electronics, 485 U.S. 717, 730 (1988) ("restraints imposed by agreements between competitors have traditionally been denominated as horizontal restraints, and those imposed by agreement between firms at different levels of distribution as vertical restraints")).

39. 290 F. 3d at 774.

40. Id. at 773.

41. Id. at 774.

42. Id.

43. Id. at 775.

44. The court twice emphasized that UST's exclusive vending relationships were not the sole basis for antitrust liability. Id. at 787 n.4, 790-91.

45. Id. at 778-79.

46. Id. at 777. Conwood produced testimony from a handful of retailers that UST representatives removed Conwood racks without permission. Deposition of Georgia Gill-Elkins (June 25, 1999) at 14-30, 53, 79-81, 96; Deposition of Marty Stevens (July 7, 1999) at 15-26; Deposition of Gayleen Rusk (May 17, 1999) at 6-29, 70, 82-83.

47. Deposition of Shawn Uluzio (Sept. 2, 1999) at 24-26, 34, 44-45, 49; Deposition of Alan Hart (Aug. 17, 1999) at 11-13; Deposition of John Naifeh (Sept. 3, 1999) at 29-33, 36-37, 42-44, 47-50. Conwood sales representatives testified that UST employees had told them that their compensation or bonuses depended on rack removal and destruction. 290 F.3d at 779-80.

48. Id. at 773.

49. Id. at 783.

50. Id. at 787 n.4.

51. See 3A Philip Areeda & Donald F. Turner, Antitrust Law, P782(a), at 272.

52. 290 F.3d at 787 n.4. Herbert Hovenkamp describes the requirement that plaintiffs disaggregate injuries caused by anticompetitive behavior from those caused violations of the antitrust laws as "well established in antitrust law, and easily stated: the damages expert must quantify the amount by which the anticompetitive conduct is responsible for the plaintiff's harm." Hovenkamp, supra note 35, at 179.

53. See, e.g., Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 600-01 (1985).

54. 290 F.3d at 787 n.4.

55. Id. at 790.

56. Id. at 784. One commentator summarizes current antitrust law as "routinely sustain[ing] the legality of exclusive dealing arrangements with foreclosure percentages of 40 percent or less." See Jonathan M. Jacobsen, Exclusive Dealing, "Foreclosure," and Consumer Harm, 70 Antitrust L.J. 311, 325, citing cases at n.85 (2002). Professor Hovenkamp suggests 20% as the minimum foreclosure percentage and 50% as a level at which courts should routinely condemn foreclosure. See XI Herbert Hovenkamp, Antitrust Law ¶ 1821c (1988).

57. 290 F.3d at 775.

58. Petition for Writ of Certiorari at 19.

59. Appellants Brief at 11.

60. See, e.g., Areeda & Hovenkamp, Antitrust Law ¶ 782a ("the antitrust court must, therefore, insist on a preliminary showing of significant and more than temporary harmful effects on competition (and not merely upon a competitor or customer) before considering a tort as an exclusionary practice. In the absence of such a preliminary showing, the defendant should win summary judgment"); Hovenkamp, supra note 35, at 178-80; accord American Professional Testing Serv., Inc. v. Harcourt Brace Jovanovich Legal & Prof'l Publ'ns, Inc., 108 F.3d 1147, 1152 (9th Cir. 1997) ("while false or misleading advertising directed solely at a single competitor may not be competition on the merits, the [conduct] in question must have a significant and enduring adverse impact on competition itself in the relevant markets to rise to the level of an antitrust violation").

61. See Deposition of Georgia Gill-Elkins (June 25, 1999) at 14-30, 53, 79-81, 96; Deposition of Marty Stevens (July 7, 1999) at 15-26; Deposition of Gayleen Rusk (May 17, 1999) at 6-29, 70, 82-83.

62. See, e.g., Testimony of Kim Overby, Trial Transcript, Vol. 16, at 13, 17-18, 24-26; Deposition of Andy Carter (March 1, 1999) at 26, 29, 41, 48, 66, 70, 145-46, 148, 166-67; Deposition of Shawn Ulizio (Sept. 2, 1999) at 27, 31-32, 38, 44, 46-47, 51, 61-62, 66; Deposition of Leroy Kyle (June 5, 1999) at 17, 42, 57-58, 64; Deposition of Alan Hart (Aug. 30, 1999) at 27-28, 47, 50; Deposition of Joseph Marconi (Aug. 17, 1999) at 12-15, 22-23, 32; Deposition of John Naifeh (Sept. 3, 1999) at 59, 63-65, 86-87, 97-98.

63. 290 F.3d at 778. Testimony of William Rosson, Trial Transcript, Vol. 2, at 95-96.

64. Testimony of William Rosson, Trial Transcript, Vol. 2, at 101; 290 F.3d at 778. Conwood did not have a rack for UST employees to remove in a substantial fraction of the stores that distributed its products. To destroy 20,000 racks per month, UST employees would have had to repeatedly remove and destroy racks from the same retailer over a significant period of time. For example, one Conwood salesperson in a large district where Conwood had been successful in promoting Kodiak testified that Conwood had racks in approximately 50-55% of his 2,000 stores. Testimony of Louis Marano, Trial Transcript, Vol. 1, at 192-193. If Conwood racks were in as many as 120,000 of stores distributing Conwood product, each store would average two instances of unauthorized product destruction or removal over the course of the year.

65. Trial Testimony of Robert Blattberg, Trial Transcript, Vol. 11-A, at 80-81.

66. This amounts to one rack per fifteen stores each month when spread over the 300,000 retail outlets estimated to be carrying moist snuff. 290 F.3d at 778. The replacement costs amount to approximately thirty-three cents per store per month. See Hovenkamp, supra note 35, at 177 (suggesting that "far from showing a significant contribution to market power, these numbers indicate that the competitive impact of rack replacement was de minimis"). Since Conwood was distributed at 81% of total outlets, or approximately 240,000 stores, spreading these costs over the 240,000 stores distributing Conwood product results in the replacement of one rack in every twelve stores per month, at a cost of approximately forty-two cents per store per month.

67. This calculation assumes a 250 day work-year.

68. Testimony of Harold Price, Trial Transcript, Vol. 6, at 267-68, 274 (discussing category management capabilities of Swisher, Swedish Match, and UST); Testimony of William Rosson, Trial Transcript, Vol. 2, at 99, 218 (conceding that Swisher, Swedish Match, and UST offer category management services, while Conwood chooses not to seek exclusive vending or offer category management). Swisher, for instance, was awarded a co-captain position for moist snuff, along with UST, at Wal-Mart. Testimony of William Rosson, Trial Transcript, Vol. 2, at 223. Wholesalers also compete with moist snuff manufacturers for the right to deliver category management services to major retail chains such as Kroger. See Testimony of Stephen R. Luckett, Trial Transcript, Vol. 11-A, at 119. While each of these suppliers and wholesalers competed for the right to become category manager with at least some success, Conwood chose not to compete. See, e.g., Testimony of William Rosson, Trial Transcript, Vol. 2, at 218; but cf. Deposition of Waynell Renfro (Sept. 28, 1999) at 33, 42-43, 51, 61, 66-67, 74-75 (testifying that Conwood gave some recommendations to Wal-Mart with respect to moist snuff category setup and product mix).

69. See supra note 66.

70. See, e.g., Deposition of Gayleen Rusk (May 17, 1999) at 70 (the store manager was successfully able to switch back removed racks after unauthorized removal and did not complain to UST); Deposition of Marti Stevens (July 7, 1999) at 43 (removed Conwood rack replaced within three days by after manager contacted Conwood); Deposition of Alan Hart (August 30, 1999) at 64-65 (testifying that UST representatives sometimes removed Conwood racks without authorization, but either UST or Conwood would replace the racks at the request of the retailer); Deposition of Patricia Leinen (July 10, 1999) at 62-65 (Conwood representative would replace removed rack whenever he came to the store and retailer never instructed UST representative not to remove racks).

71. See, e.g., D.H. Kaye, The Dynamics of Daubert Methodology, Conclusions, and Statistical Fit in Econometric Studies 87 U. Va. L. Rev. 1933 (2001); D.H. Kaye, Adversarial Econometrics in United States Tobacco v. Conwood Co., 43 Jurimetrics 343 (2003); Hovenkamp, supra note 35, at 81-91; Brief of Washington Legal Foundation, Stephen E. Fienberg, Franklin M. Fisher, Daniel L. McFadden, and Daniel L. Rubinfeld as Amici Curiae in Support of Petitioners (No. 02-603).

72. Hovenkamp, supra note 35, at 86-88, summarizes and criticizes this testimony.

73. 290 F.3d at 793. As Kaye points out, supra note 71, the expert actually compared Conwood's share growth rates in "low" share states with a hypothetical state in which Conwood had a 20% share in 1990. Kaye demonstrates that a comparison of share growth in high- and low-share states reveals no statistically significant difference. Kaye, supra note 71, at 2003 n.99. Hovenkamp, supra note 35, at 81, also reports that removing Washington D.C. from the expert's analysis as an outlier would render the relationship between Conwood's starting share and subsequent growth insignificantly different from zero.

74. See, e.g., Kaye, supra note 71, at 3451-52 ("the Sixth Circuit refused to accept an informative amicus brief outlining the requirements for a valid econometric analysis of damages and explaining why Conwood's study did not measure up to these standards"); Hovenkamp, supra note 35, at 88 ("the methodology flunked all of the criteria the Supreme Court provided in Daubert"); Brief as Amici Curiae, supra note 71, at 9 (describing the regression analysis as "unacceptable science").

75. Id. at 775 n.1. Swisher won similar competitions for exclusive racks at K-Mart and Tom Thumb stores. Id.

76. 290 F.3d at 778.

77. Other manufacturers offered similar discount programs, including Swisher and Swedish Match. Appellants Brief at 24.

78. Testimony of William Rosson, Trial Transcript, Vol. 2, at 218.