On September 9, 2004, the United States District Court for the Northern District of California, Judge Vaughn Walker, issued its much anticipated opinion in the Department of Justice’s ("DOJ"), case against Oracle Corp.’s ("Oracle") attempted acquisition of PeopleSoft, Inc. ("PeopleSoft"). Oracle and PeopleSoft both develop, manufacture, and sell "enterprise resource planning" ("ERP") software, designed to "integrate[] most of an entity’s data across all or most of the entity’s activities."1 In more basic terms, ERP software is used by a company to help manage specific business/administrative functions, such as human relations management ("HRM"), financial management systems ("FMS"), customer relations management ("CRM"), or supply chain management ("SCM").

The theory behind the DOJ’s case was that Oracle and PeopleSoft are two of the three largest producers of HRM and FMS ERP software for large businesses in the U.S. In each of the U.S. markets for "high function" HRM and FMS software, the DOJ alleged, there are only three competitors – Oracle, PeopleSoft and SAPAG, a German software company with operations in the U.S. According to the complaint, if allowed, Oracle would acquire its closest rival, conferring on it the ability to raise prices unilaterally.2

The District Court found in favor of Oracle and denied the DOJ’s request for an injunction. In a 36-page published opinion that is highly fact-intensive and relies heavily on antitrust economics, the Court detailed its disagreement with virtually every DOJ allegation. The Court completely rejected the alleged product market definition of "high function" HRM and FMS software targeted to large businesses (a product market that included only Oracle, PeopleSoft, and SAP), finding instead that software companies that make ERP software for "mid-market" customers also compete for those same "large complex enterprises." The Court also found that other solutions competed with HRM and FMS ERP software, namely outsourcing and best-of-breed solutions.3 The Court also disagreed with a geographic market definition limited to the U.S.4 Finally, even assuming these market definitions as correct, the Court found that there was no localized competition between Oracle and PeopleSoft, the elimination of which could possibly lead to higher prices through unilateral effects.5

The Court’s opinion is of particular interest not just because of its detailed analysis of a unilateral effects case, but also because of what it teaches us about the types of evidence a court is likely to find persuasive (or in this case, unpersuasive) in proving the basic elements of a merger case. Three of the more important lessons are:

1. The Product Market Definition Must Be Supported by Industry Realities

Much of the Court’s opinion was spent dispelling the DOJ’s proposed market definition – "high function HRM software" and "high function FMS" software for use by "large complex enterprises." These products, the DOJ alleged, compete in markets separate and distinct from other ERP products, like CRM or SCM software, and mid-market HRM and FMS software, products designed for smaller companies that have much less demanding needs.

The Court, however, was able to point to numerous examples of documents and other industry evidence refuting such a narrow market definition. For example, the terminology "high function software" has no recognized meaning in the industry. Also, there was no bright line test for what constitutes a large or up-market customer versus a smaller, mid-market customer. For instance, PeopleSoft considered large customers as those with more than $500 million in revenue, while SAP’s definition included companies with revenues of more than $1.5 billion. Oracle’s definition did not depend on revenues at all, but the number of employees. To make matters worse, the DOJ tried to distinguish between mid-market and large customers based on the amount of money spent on ERP purchases, a measure that was both unreliable and not substantiated by any industry facts. Additionally, the Court found that the ERP software sold to large customers is the same as that sold to mid-market customers.6

Particularly harmful to the DOJ’s case were its industry witnesses who were effectively impeached on the issue of market definition. For example, one witness from PeopleSoft, in trying to limit the market to Oracle, PeopleSoft, and SAP, testified that Lawson (an ERP vendor the DOJ labeled as mid-market) never competed with PeopleSoft for large customers.7 On cross examination, however, Oracle produced a document showing that Lawson had competed with PeopleSoft on 27 occasions. Other witnesses also were contradicted by documents or other evidence concerning the competitiveness of Lawson for the business of large customers, leading the Court to dub the phenomenon "Lawson Amnesia."8

Perhaps the most embarrassing piece of evidence, though, came from DOJ’s own response to market conditions. DOJ itself had chosen another ERP vendor, AMS – a vendor that the DOJ claimed was a mid-market vendor and not in the relevant market – to provide FMS software over the offerings of Oracle, PeopleSoft, and SAP.9

2. Customer Testimonials May Be Insufficient Without Supporting Economic Evidence

To prove the existence of its "high function" HRM and FMS markets, the DOJ offered the testimony of ten customers of ERP software, all falling within the DOJ’s definition of large customers. Major organizations such as DaimlerChrysler, the State of North Dakota, Pepsi Americas, Nextel, Greyhound Lines, and Verizon testified in support of the notion that their choices for ERP software were limited. Eight of the customer witnesses testified that only Oracle, PeopleSoft, and SAP offered HRM and FMS solutions that met their needs. More importantly, eight of these customers also testified to what is generally considered the determinative factor in defining the product market – that they would not consider any manufacturers of HRM and FMS software other than Oracle, PeopleSoft, or SAP, or any other ERP solutions (such as outsourcing or best of breed vendors), even if the price of HRM and FMS software increased by ten percent.

But the Court quickly dismissed this testimony as "largely unhelpful to plaintiffs’ effort to define a narrow market of high function FMS and HRM."10 Focusing on the du Pont standard of reasonable interchangeability, the Court found that the DOJ failed to show what customers could do in the face of a price increase, and instead only showed what these ten customers would prefer to do. "Customer preferences," the Court stated, "towards one product over another do not negate interchangeability."11 The Court found no evidence of what customers could do to avoid a price increase. Nor did any of the DOJ’s customer witnesses substantiate their claims that they would not switch to other solutions with any "cost/benefit analyses of the type that surely they employ and would employ in assessing an ERP purchase."12

In contrast, Oracle presented two customer witnesses – both considered large customers under the DOJ’s definition – who testified that they had turned to other options, including other ERP vendors, outsourcing, and best-of-breed solutions, instead of the products offered by Oracle and PeopleSoft. The Court found these witnesses more creditable than the DOJ’s witnesses since they were able to show "concrete and specific actions" that they had taken to meet their company’s needs.13

3. Evidence of Head-to-Head Competition between the Merging Parties Does Not Spell the Doom of the Transaction

To prove its case of unilateral effects, the DOJ presented evidence that Oracle and PeopleSoft competed aggressively head-to-head and were each other’s closest substitutes. The DOJ entered into evidence several quarterly "win/loss analysis" documents comparing Oracle’s loss percentage to PeopleSoft versus its loss percentage to SAP. These win/loss analyses showed that Oracle lost to PeopleSoft more often than it lost to SAP, which the DOJ claimed showed that Oracle and PeopleSoft were each other’s closest competitors. In fact, at least one Oracle document described PeopleSoft as its "Number #1 competitor," and SAP as its "Number #2 competitor."14

The DOJ also offered expert testimony analyzing discounts offered by Oracle when PeopleSoft competed for the same business. According to the DOJ’s expert, discount documents showed large Oracle discounts when PeopleSoft was competing for the same business. A regression analysis showed that Oracle’s discounts were 9 to 14 percent higher when PeopleSoft was in competition than when PeopleSoft was not competing for the business.15

The Court, however, did not find this evidence of localized competition persuasive. As for the win/loss documents, some of the same documents relied on by the DOJ showed that Oracle’s loss percentage to PeopleSoft was only slightly larger than its loss percentage to SAP. One document, for example, showed Oracle losing to PeopleSoft 54 percent of the time while also losing to SAP 53 percent of the time – a mere 1 percent difference.16

Also troubling to the Court was the fact that even the DOJ’s own expert credited SAP with considerable competitive presence in the market. Assuming the DOJ’s market definitions as true, SAP had the largest market share (39 percent) in the high function FMS market and the second highest share (29 percent) behind PeopleSoft in the high function HRM market. (Oracle had the smallest share of the three in both markets.) This suggested that SAP was beating Oracle in both the FMS and HRM markets and PeopleSoft in the FMS market. The Court found it hard to believe that SAP was somehow disadvantaged in competing against Oracle and PeopleSoft in a way that would allow the post-acquisition Oracle to raise prices unilaterally.17

As for the expert testimony, the Court recognized that Oracle and PeopleSoft were fierce competitors and that when they met head-to-head, they drove down each other’s prices. What this had to do with "whether Oracle and PeopleSoft are competing head to head in a product space in which SAP is not a party" escaped the Court.18 The discount documents and the regression analyses all showed that Oracle offered large discounts when PeopleSoft was in competition with it, but there was no showing of the discount level (or presumably lack of discount level if the DOJ’s theory was correct) when Oracle competed with SAP. Central to any unilateral effects case, the Court found, was that not only are the parties close substitutes, but that other products are sufficiently different so as not to constrain the prices of the post-merger firm. As the Court stated, "[s]imply because Oracle and PeopleSoft often meet on the battlefield and fight aggressively does not lead to the conclusion that they do so in the absence of SAP."19 Without evidence that SAP did not provide effective price competition to Oracle, the DOJ could not prove the type of localized competition between the merging parties that would give rise to a unilateral effects case.

Conclusion

The Oracle case provides significant insights into the types of evidence the federal antitrust agencies must present to block a transaction. Not since the FTC’s victory in FTC v. Staples have we seen such a detailed and instructive opinion on the law of merger review. Of all the lessons that can be pulled from the opinion, three predominate – 1) the proposed market definition must be consistent with and supported by industry facts; 2) the touchstone of customer testimony about what they would prefer must be fully tested against what they economically could do to avoid a price increase; and 3) evidence of head-to-head competition between the parties is insufficient to prove unilateral effects; there must also be a lack of competition from other firms.20

Endnotes

1 U.S. v. Oracle Corporation, 331 F. Supp. 2d 1098, 1101 (N.D. Cal. 2004).

2 The DOJ did not pursue a theory of coordinated interaction, wherein the elimination of a competitor would increase the likelihood that the remaining competitors could tacitly coordinate their competitive conduct.

3 "Outsourcing occurs when a company hires another firm to perform business functions, often HRM functions." Oracle, at 1106. Best of breed vendors focus their product offerings on one type of ERP application, or "pillar," such as customer relations management ("CRM"). Id. Companies like Oracle or PeopleSoft sell many different pillars. Their HRM and FMS pillars typically are sold in a bundle, in which the vendor offers a discount for taking the entire bundle. Customers can force companies like Oracle or PeopleSoft to offer larger discounts by threatening to use a best-of-breed vendor for certain pillars instead of taking the entire bundle. Id.

4 Oracle, at 1108.

5 Oracle, at 1109.

6 Oracle, at 1106.

7 Because this was a hostile takeover by Oracle, the PeopleSoft executives were testifying on behalf of the DOJ.

8 Oracle, at 1139.

9 Oracle, at 1105.

10 Oracle, at 1130.

11 Oracle, at 1131.

12 Id.

13 Oracle, at 1133.

14 Oracle, at 1166.

15 Oracle, at 1169.

16 Oracle, at 1166.

17 Oracle, at 1167.

18 Oracle, at 1169.

19 Id.

20 The DOJ did not appeal the decision and Oracle completed the acquisition in December, 2004.

Copyright © 2007, Mayer, Brown, Rowe & Maw LLP. and/or Mayer Brown International LLP. This Mayer Brown article provides information and comments on legal issues and developments of interest. The foregoing is not a comprehensive treatment of the subject matter covered and is not intended to provide legal advice. Readers should seek specific legal advice before taking any action with respect to the matters discussed herein.

Mayer Brown is a combination of two limited liability partnerships: one named Mayer Brown LLP, established in Illinois, USA; and one named Mayer Brown International LLP, incorporated in England.