United States: Mind Your Markets
Arguments in Oracle case highlight ongoing debate over market definition.

Previously published in Legal Times, July 2004

Antitrust challenges to mergers very often turn on how the enforcement agencies and the courts define the "relevant markets" affected by the merger. Market analysis entails a highly fact-specific inquiry into various economic factors, yet still remains as much an art as a science. The difficulties with market analysis have recently taken center stage in the government’s attempt to block the Oracle Corp.’s hostile bid for PeopleSoft Inc.

The Justice Department’s Antitrust Division alleges that the merger would make a concentrated business software market even more concentrated, taking the market from three down to two competitors. Oracle contends that the DOJ has gerrymandered the relevant market, unduly restricting the products supposedly competing in the market and improperly limiting its geographic dimension.

This battle over market definition may largely determine whether the U.S. District Court in San Francisco, which is now hearing the case, blocks the acquisition. The market definitions put forward by each side seem custom-fit to their subjective arguments. Antitrust law has faced this clash before.

Antitrust merger analysis requires identification of the "relevant market" that the merger may affect. A relevant market has two components— product and geography. The relevant product market includes all products that purchasers consider reasonable substitutes for the product in question. The relevant geographic market covers the area in which purchasers are reasonably likely to seek out sellers. The federal antitrust agencies have set out the paradigms used in defining markets in their Horizontal Merger Guidelines. Courts have recognized these guidelines as acceptable, although not necessarily binding on them. Both paradigms, for product and geographic markets, have some degree of subjectivity.

The test for defining the product market under the guidelines assumes a hypothetical monopolist that controls all potentially competing products, trying to impose a price increase—usually 5 percent— while prices for other products remain unchanged. If customers would find and switch to alternative products in sufficient numbers to make the increase unprofitable, these alternative products should be included in the market.

The guidelines use essentially the same paradigm for the geographic market. If purchasers can effectively turn to products sold outside the borders set in the hypothesis in order to avoid the price increase, the geographic scope of the market should be expanded to include the areas in which those other products are sold.

In the Oracle case, both the product and geographic market definitions play a critical role. Indeed, Oracle suggests that if the government fails to prevail on either point, the concentration levels and market shares in the post-acquisition market will be insufficiently high as to warrant competitive concerns.

The DOJ argues that relevant markets exist for "high-function" enterprise software for both human resources and financial management applications, and that competition is limited to the United States. According to the government, high-function software is software that can execute a wide array of business processes at a superior level of performance. The government alleges that only three companies— Oracle, PeopleSoft, and the German software company SAP— compete in these relevant markets and that "mid-market" software is not competitive.

As alleged in the DOJ’s trial brief, mid-market software—a term that is undefined in the joint submission of definitions—does not compete against high-function software because mid-market products offer only basic human resources and financial management functionality and, unlike high-function products, cannot "handle large data sets, large numbers of transactions and concurrent users, and multiple languages and currencies." To bolster its position, the DOJ alleges that Oracle often discounts against PeopleSoft and SAP but not against others—a fact supposedly borne out by Oracle’s documents. This would seem to support the government’s product market definition.

But Oracle scorns the government’s market definition as completely artificial, calling it the type of "strange, red-haired, bearded man with-a-limp" market that courts have derided in the past. In particular, the distinction between high-function and mid-market software, according to Oracle, is false. The company argues that several multibillion-dollar enterprises and the DOJ itself—the very types of organizations that supposedly need high-function software—have bought mid-market products. Customers or vendors can purchase additional modules and "stitch" them into the mid-market software to achieve the same functionality as high-function products, says Oracle. Thus, Oracle asserts that the government fails to take into account several viable alternatives to which customers can and already do turn.

Whether differences in grade and quality can justify separate markets for products having similar functions frequently comes up in antitrust cases. And courts have come down on both sides of this issue. Courts have identified distinct markets for first-run major motion pictures, championship boxing matches, and high-quality boat anchors. Other courts have refused to divide markets along quality lines when asked to draw distinctions between grades of ice cream, furniture, and motorcycle valves.

In the past, the antitrust enforcement agencies have argued market definitions based on quality and grade differences. In 1993, the DOJ sought to block a merger in what the agency referred to as the premium fountain pen market. The court accepted the concept of a premium market but included all premium writing instruments, e.g., mechanical pencils. In that broader market, the merger between the Gillette Co. and Parker Pen Holdings posed no threat.

More recently, in 2003, the Federal Trade Commission sued to block a merger in the ice cream industry, between Nestlé Holdings and Dreyer’s Grand Ice Cream, on the basis of alleged harm in the super-premium ice cream market. Interestingly, the FTC’s proposed market directly conflicted with earlier cases dealing with the very same products, which found that a super-premium ice cream market did not exist. The FTC settled the case without having to litigate the market definition issue.

Oracle also challenges the government’s "one stop shopping" argument—i.e., that customers prefer to purchase software with all the bells and whistles from a single vendor. Oracle points out that the U.S. Court of Appeals for the 9th Circuit, which will hear any appeal in the case, rejected a one-stop-shopping argument in 1989 in connection with home improvement products sold through what have commonly become known as superstores.

The FTC, however, successfully used the one-stop-shopping argument in challenging the Staples/Office Depot merger in 1997. In that case, the FTC’s definition of the product market—sales of office supplies through superstores—raised eyebrows because office supplies are available through many sources, and therefore many people had difficulty believing that office supply superstores could charge appreciably higher prices than other outlets. But the FTC showed that the merging parties "believed" that competition effectively took place only between superstores, and that prices charged by superstores were higher in those geographic markets where no other superstores were present. Based on this evidence, the court defined a relevant market limited to superstores and enjoined the merger. Similarly, the DOJ in the Oracle case claims that pricing data show that competition from mid-market vendors does not constrain the prices for high-function software.

The geographic market definition for Oracle and PeopleSoft has proved similarly contentious. According to the DOJ, the geographic market is limited to the United States, apparently because prices overseas do not affect prices domestically and vice versa.

Oracle disagrees. The proper geographic scope of the market is worldwide, it argues. SAP operates out of Germany, and Oracle and PeopleSoft both sell their software internationally. Oracle makes a rather compelling point that, even under the government’s proposed product market definition, software from outside the United States, specifically from SAP, could be imported to thwart a price increase by Oracle and PeopleSoft. Case law and the federal merger guidelines would seem to suggest a geographic market broader than merely the United States.

Moreover, the government’s allegations regarding the product market may undermine its proposed geographic market. The DOJ alleges that large organizations with multinational operations need high-function software. As Oracle points out, there is a certain inconsistency in saying that multinational enterprises buy the product to use in their offices throughout the world, yet the market stops at the U.S. border. Oracle and the government each raise legitimate arguments to support their proposed market definitions, and each can point to case law seemingly buttressing their arguments. Ultimately, however, when it comes to defining a relevant market, legal precedent usually takes a back seat to the facts of the specific case. In Brown Shoe Co. v. United States (1962), the Supreme Court noted that "Congress prescribed a pragmatic factual approach to the definition of the relevant market and not a formal, legalistic one."

Whatever relevant markets the court in the Oracle case may select must correspond to the commercial realities of the industry. Accordingly, the particular evidence presented by the parties, the credibility of that evidence, and the weight the court gives to that evidence will ultimately drive the relevant market determination.

Because market definition depends so heavily on the particular facts, it is not clear that the outcome in the Oracle case, whatever it may be, will have a measurable effect on future merger challenges. Nonetheless, Oracle raises some interesting questions about how the government approaches the relevant market inquiry in software and other high-tech mergers. Generally, the trend in software cases, which mostly have not been litigated, has been for the government to argue for very narrow product market definitions, limited, for instance, to software capable of performing one or perhaps a handful of fairly specific and finite types of functions, e.g., debugging software.

The DOJ’s approach against Oracle seems to deviate from this trend. Both the human resources and financial management software in question contain somewhere between 30 and 70 different modules, which apparently can be added to and segregated from the overall software package. By seeking product market definitions for bundles of software programs, the DOJ has arguably described a market having much greater breadth than in previous software cases.

Whether that portends a new trend remains to be seen. Indeed, we don’t even know yet whether the argument will work against Oracle.

The information contained in this article is not intended as legal advice or as an opinion on specific facts. For more information about these issues, please contact the author(s) of this article or your existing firm contact. The invitation to contact the author is not to be construed as a solicitation for legal work in any jurisdiction in which the author is not admitted to practice. There will be no charge for the initial contact. Any attorney/client relationship must be confirmed in writing. You may also contact us through our Web site at www.kilpatrickstockton.com