Eighteen months after the deal was first announced, Sysco Corporation (“Sysco”) and US Foods, Inc. (“USF”) abandoned their $3.5 billion merger following the Federal Trade Commission’s (the “FTC” or “Commission”) decisive victory in obtaining a preliminary injunction blocking the transaction. FTC v. Sysco Corp., 1:15-cv-00256 (D.D.C. June 26, 2015). While Judge Mehta’s comprehensive opinion can in large measure serve as a hornbook for judicial merger analysis, the decision is significant in its acceptance of the FTC’s view that a particular customer type can serve as the basis for defining the relevant antitrust market. The decision also provides valuable insight as to the analysis a court might apply to a proposed divestiture remedy designed to adequately restore competition and thus salvage an otherwise anticompetitive deal.
Acknowledging that the parties had announced their intent to abandon the transaction if a preliminary injunction was granted, the district court recognized the “real-world impact” of its decision beyond the question presented of whether the transaction should be enjoined until it could be fully reviewed on the merits by an FTC Administrative Law Judge. The district court’s opinion on the FTC’s Section 13(b) motion for a preliminary injunction reads as a full decision on the merits, focused largely on the FTC’s market definitions. While merger cases brought by the antitrust agencies frequently heavily rely on evidence of price competition between the parties and so-called “bad documents” (and documents did play a confirming role in this case), the FTC here emphasized the presumption of anticompetitive effects based on market shares and market concentration derived from its market definitions.
The case is one of the first major decisions for U.S. District Court Judge Mehta who was just confirmed last December. The FTC investigated the proposed merger, announced by the parties in December 2013, for more than a year before voting 3-2 to file an administrative complaint in February 2015 to block the transaction. The FTC sought a temporary restraining order and a preliminary injunction in federal court to prevent the parties from consummating the deal prior to the conclusion of the administrative proceeding on the merits. Over two short months, the case included millions of discovery documents, dozens of depositions, nearly 100 declarations by competitors and customers, and eight days of testimony, which included dueling economic testimony.
The Food Service Distribution Industry
While the decision goes into great detail about the complex food service distribution industry, a review of at least the basics is necessary to understand the district court’s analysis and findings with regard to the antitrust relevant markets here.
The $230+ billion foodservice distribution business supplies food and related products to restaurants, hospitals, cafeterias, sports arenas, and a host of other places that serve prepared food. As many as 16,000 companies compete at some level in the foodservice distribution marketplace, with Sysco and USF being the two largest. The industry generally recognizes four categories of foodservice distribution companies: (i) broadline distributors, (ii) systems distributors, (iii) specialty distributors, and (iv) cash-and-carry or club stores. Broadline distributors are characterized by their product breadth, private-label product offerings, frequent/flexible delivery, and value-added services. Systems distributors carry and deliver proprietary products manufactured for specific customers. Specialty distributors offer and deliver limited product categories, such as meat or dairy or seafood. Finally, cash-and carry stores offer a self-service model of food distribution.
Within the broadline category, there are three types. National broadliners have their own nationwide service capabilities. There are only two such companies: Sysco and USF. Regional broadliners generally offer the same products and services as national broadliners, but their distribution capabilities are concentrated in discrete regions of the country. The largest regional broadliner is Performance Food Group, Inc. (“PFG”). There are also many local broadline companies that operate within more limited geographic areas. Several regional broadliners have created a joint venture that collectively provides nationwide coverage to compete for accounts looking for nationwide service.
Foodservice distribution customers may purchase from a single distribution channel or through a mix of all four. Customers range from large group purchasing organizations to small “street” customers such as independent restaurants.
District Court’s Analysis
The FTC’s complaint challenged the proposed merger as a violation of Section 7 of the Clayton Act. Section 13(b) of the FTC Act allows the Commission to seek a preliminary injunction to block a merger pending the administrative trial if such injunction would be in the public interest. Under this public interest standard, the FTC need only show that there is a reasonable probability that, after the administrative trial on the merits, it will succeed in proving that the proposed merger may substantially lessen competition or tend to create a monopoly. Although the public interest standard is more relaxed than a traditional equity injunction standard, the district court here recognized that the issuance of a preliminary injunction is an “extraordinary and drastic remedy.”
The district court noted that “[m]arket definition [was] the parties’ primary battlefield in this case.” The FTC alleged two relevant product markets: i) a product market for broadline distribution; and ii) a sub-product market for broadline distribution to national customers. The Commission’s alleged relevant geographic markets were local for local customers and national for national customers. The parties countered that the food distribution market “cannot be sliced and diced” and limited to only broadline distributors, but rather the relevant product market should include all food distribution channels.
Relevant Product Market
Although Judge Mehta used the Merger Guidelines in other parts of his analysis, he grounded his market definition holdings on Supreme Court precedent. Applying the Supreme Court’s “reasonable interchangeability of use or the cross-elasticity of demand” standard for defining the relevant product market, the district court here ultimately agreed with the FTC. Brown Shoe Co. v. United States, 370 U.S. 294, 325 (1962). On the issue of whether “the product market should be expanded to include other modes of distribution,” the district court found that they should not be included because several characteristics of broadline distribution fulfill the “practical indicia” test from Brown Shoe, including: i) the breadth and diversity of products offered; ii) the size and scale of the facilities and operations; iii) the timeliness and flexibility of delivery; iv) the availability of value-added services; v) the types of customers capable of being served; vi) pricing competition only against other broadline distributors; and vii) industry recognition of broadline distribution as distinct from other modes of distribution. While not disputing these distinguishing characteristics of broadline distribution, the merging parties argued that other channels of distribution compete and constrain price, citing independent restaurants and fast food chains as examples of customers that buy across multiple distribution channels. The district court, however, held that “the fact that Defendants sometimes compete against other channels of distribution in the larger marketplace does not mean that those alternative channels belong in the relevant product market for purposes of merger analysis.” (The district court relied on its own precedent from FTC v. Staples, Inc., 970 F. Supp. 1066 (D.D.C. 1997), as well as FTC v. Whole Foods Market, Inc., 548 F.3d 1028 (D.C. Cir. 2008), explaining that “fruit can be bought from both a grocery store and a fruit stand, but no one would reasonably assert that buying all of one’s groceries from a fruit stand is a reasonable substitute for buying from a grocery store.”)
Central to the FTC’s argument for a national broadline distribution sub-product market was whether there is a “national customer,” which the parties strongly disputed. Here, the FTC defined a relevant product market based on a specific type of customer. The district court determined that it “need not resolve the Whole Foods disagreement over defining a market around a “core” customer.”1 Instead, it again found that the Brown Shoe practical indicia test and expert testimony supported the FTC’s position. Most compelling to the district court was the fact that regional broadliners had formed cooperatives specifically to compete for national customers, and a consultant hired by Sysco recognized a separate category for national customers, as did the parties’ ordinary course business documents. Additionally, although the district court agreed with some of the critique of the FTC expert’s relevant market analysis, it ultimately determined that the FTC expert’s conclusion was “more consistent with the business realities of the food distribution market.”
Relevant Geographic Market
The district court again relied on Supreme Court precedent to assess the relevant geographic markets, asking “where, within the area of competitive overlap, the effect of the merger on competition will be direct and immediate” (quoting United States v. Phila. Nat’l Bank, 374 U.S. 321, 357 (1963)). Having accepted the FTC’s two broadline distribution product markets, the district court also agreed with the FTC’s national and local geographic market definitions. In its argument, the FTC cited the parties’ “national account teams” dedicated to national customers, contending that for national customers the relevant geographic market is nationwide. The district court agreed, noting that the parties compete “by touting their nationwide distribution capabilities to [national customers]; bidding against other broadliners with multi-regional capabilities…; coordinating the marketing, negotiating, and managing of these customers through their ‘national account’ teams; and… pricing [that applies] across regions.”
The district court next commented that “[d]efining the local geographic market presents a far greater challenge… [because] there is no industry standard… [and] [e]ach local market has its own unique attributes.” For local customers, the FTC’s expert constructed relevant geographic markets based on proximity to the parties’ distribution centers, arguing that for local customers the distance to a distribution center is a key service factor. The FTC ultimately identified 32 local geographic markets where the parties’ combined market share would be dominant. The Commission’s expert constructed the local geographic markets as follows: first, he drew circles around each of the parties’ distribution centers using a radius based on the area covered by 75% of the center’s sales to local customers (the “draw radius”); second, he identified the local customers that fell within the overlap area between the circles drawn around Sysco and USF distribution centers (the “overlap customers”); third, he identified the other broadline distributors who could compete for the overlap customers by applying the draw radius around each overlap customer. The parties argued that this methodology was “arbitrary” and not reflective of industry realities, citing competitor declarations that showed there are competitors who are willing to and do drive distances greater than the draw radius. The district court, however, found that the FTC expert’s methodology was not arbitrary, and instead “provides a practical approach and solution to an otherwise thorny problem.”
Effects on Competition
Market Share and Market Concentration
As is frequently the case in merger cases, the experts for the parties and for the FTC disagreed on the appropriate method for calculating market shares and market concentration. The FTC’s expert asserted combined market shares of at least 59% for the national broadline customer market, and as high as 100% in some local broadline markets. The parties’ expert countered that those calculations were artificially high due to the use of incorrect inputs. (In contrast, using the parties’ product market definition that included all food distribution channels, the combined market share would be 25%.) The district court again agreed with the FTC’s expert, thus finding that the FTC had made a prima facie case and established a rebuttable presumption that the merger would lessen competition in both the local and national markets.
For market concentration in the national broadline customer market, the FTC’s expert used multiple variants for his calculations. The most convincing to the district court was a method that “almost certainly underestimated Defendants’ market shares,” but still resulted in 59%. In this method, the FTC’s expert used a numerator based on individual sales to customers identified by the parties’ “national” customer designation, and a denominator that aggregated national sales reported by the largest 16 broadliners in response to the Commission’s civil investigative demands (including consortiums of regional boardliners) plus an estimate of national sales for all other responding distributors using the assumption that each distributor’s national-local sales ratio was the same as the parties’ ratio. In crediting this method, the district court noted that the smaller broadliners were actually unlikely to have a national-local sales ratio similar to the parties, thus this method would underestimate the parties’ market share.
For local market concentration, the FTC’s expert calculated shares with the 75% draw radius used to identify the relevant geographic markets, as well as alternatives using a 90% draw radius and a 95% weighted draw radius to include more distant competitors. For each overlap customer, he then calculated the market shares for the competitors within the draw radius, and then aggregated those customer-specific shares to the local level using weighted averages across all overlap customers. The calculations were based alternatively on square footage of distribution centers, local broadline sales, and number of sales representatives. The parties criticized this methodology as unreliable because it assumed that no competitor would drive a greater distance than Sysco or USF currently do to provide broadline services. The district court agreed with the critique that the FTC expert’s methodology resulted in overstated market shares for the parties when using the 75% draw radius, but found that the alternate calculations using the 90% and 95% draw radii “provide[d] persuasive evidence of the merger’s impact on local markets.”
Additional Evidence of Competitive Harm
The FTC further bolstered its case with evidence of unilateral effects for national customers from the elimination of head-to-head competition, as demonstrated by RFP bid and win data, ordinary course business documents, and testimonial evidence. The FTC’s expert found that USF appeared as a competitor to Sysco in RFP bids for national broadline business twice as often as the next competitor, and that when Sysco lost, it lost to USF two and one half times more often than it lost to the next competitor. USF’s bid data showed similar results. Additionally, both parties’ ordinary course business documents identified each other as the “major rival” or the “next largest competitor.” Finally, testimonial evidence from industry actors also characterized Sysco and USF as close competitors for national customers. The FTC also demonstrated that the merger would harm national customers using an “auction model” which predicts that post-merger competitive pricing pressure on the merged entity will come from what had been the pre-merger third-place bidder. The auction model suggests harm to customers if, as was the case here, the third bidder was distant in terms of pricing pressure. Again, the parties’ business documents provided support for this argument.
The FTC’s expert used Sysco and USF internal databases to similarly demonstrate unilateral effects for local customers. The parties argued that the databases were never intended as a win-loss record, and thus not reliably used for that purpose. While the district court agreed that using the databases as such was problematic, it determined that it could not disregard the evidence because it “so overwhelmingly demonstrated primary competition between Sysco and USF.”
Court Finds Parties’ Proposed Divestiture Fix Inadequate
In an attempt to resolve the FTC’s concerns with the transaction, the parties had announced in early February 2015 their intent to divest 11 USF food distribution facilities to PFG—the next largest broadline distributor—if the merger received regulatory approval. This planned divestiture was a key argument in the parties’ rebuttal against the finding of a prima facie case. The district court, however, was not convinced that the divestiture would “maintain—certainly not in the near term—the pre-merger level of competition that characterizes the present market place.” In particular, PFG’s business plan showed that within five years it would still not be as competitive as USF is today. Also, PFG’s internal strategy documents indicated that it did not initially believe 11 distribution centers would be sufficient for it to compete on a national level and instead sought 15 facilities in the divestiture package.
The district court also discounted the parties’ arguments that other existing competition would constrain price increases, that entry would be timely and/or sufficient, and that the anticipated efficiencies were merger specific and sufficient to offset the likely harm to competition. Relying on the same reasoning it used to find that national broadline distribution is a relevant product market, the district court found that regional broadline distributors could not sufficiently constrain prices.
The district court stated in conclusion that “there can be little doubt that the acquisition of the second largest firm in the market by the largest firm in the market will tend to harm competition in that market.” Ultimately, despite the lengthy review of the extensive evidence and expert debates in the 128-page opinion, the FTC won this case on market definitions. Although the district court did not resolve the “core customer” debate, its holdings here suggest that a market can be defined based on a particular type of customer’s need. Furthermore, while there was no “smoking gun” document here, the district court repeatedly pointed to evidence from the parties’ ordinary course documents that supported the FTC’s position. Finally, the court did not hold the FTC’s expert to a standard of water-tight perfection, but found the testimony sufficient even though the defendants had successfully exposed certain flaws in the data and approach.
This decision also highlights the importance of presenting a divestiture remedy constructed to immediately replace lost competition, not just maintain the number of competitors. If the proposed remedy does not satisfy the agency during the investigation stage, it is unlikely to save a merger in court once a rebuttable presumption of anticompetitive effects has been made. Of note, the lead attorney for the FTC recently said that the parties’ proposed divestiture here “had value because it showed the court and others interested in the matter that ... there was some validity to our concerns.”
1 The United States Court of Appeals decision enjoining the Whole Foods/Wild Oats merger held that “core consumers can, in appropriate circumstances, be worthy of antitrust protection” (citing the Merger Guidelines which explain the possibility of price discrimination for “targeted buyers”). The Court of Appeals explained that the district court erred in ignoring evidence that Whole Foods and Wild Oats competed for core consumers in a premium market even if they also competed on individual products for marginal consumers in a broader market.