United States: Federal Merger Enforcement: Is There A Broader Message In Recent Losses?

In what has become the highly regulatory world of Hart-Scott- Rodino federal merger review, it is easy to forget that this is still law enforcement — the Federal Trade Commission or the Antitrust Division cannot block any merger by administrative fiat. Instead, they must go to federal district court and carry their burden of proof before an independent federal judge. For the last several decades, that burden has been significantly eased by the presence of various judicial or statutory presumptions. As a result, for most of that period the agencies have been generally quite successful in merger litigation (if we leave aside hospital mergers, where their record is dismal).

Among the most important of these presumptions is the Philadelphia National Bank-created presumption that increases Bank in market concentration are a predictor of anticompetitive effects. Another, not really a presumption but with a similar practical effect, is the FTC Act’s Section 13(b) statutory language that all the FTC must show to be entitled to a preliminary injunction is that there are "substantial, difficult and doubtful" questions going to the merits; many courts treated this as requiring not much more than a prima facie case, and sometimes not even that. These and other similar analytical shortcuts and legal crutches not surprisingly led both agencies to generally offer a more simplistic, formalistic analysis in their court pleadings and evidence than they in fact applied in making the decision whether to file the lawsuit in the first place. Most of the time, this relatively simple approach has been sufficient to obtain the preliminary injunction sought and, again most of the time, that was enough to kill the deal.

To a significant extent, the concept of merger enforcement by presumptions can be traced to a seminal law review article in 1960 by Derek Bok, then a law professor and subsequently Dean of Harvard Law School and President of Harvard University. [D. Bok, Section 7 of the Clayton Act and the Merging of Law and Economics, 74 Harv. L. Rev. 226 (1960)]. Bok, writing not that long after the 1950 Celler- Kefauver amendments to the Clayton Act, argued that it would be administratively impossible to carefully analyze every merger for its actual anticompetitive potential, and in any event such a process would not produce useful guidance for counselors or the business community. Thus, he suggested the creation of relatively bright-line rules, based on economic learning and experience, that could guide the enforcement of what was then a reinvigorated Section 7. Philadelphia National Bank followed three years later, obviously influenced by this approach, and the very first DOJ Merger Guidelines came out in 1968, setting forth fairly bright-line standards that were very influential in the courts.

While this general approach to merger analysis has survived for more than four decades, over time we have seen a steady and increasing trend toward more detailed analysis and weaker presumptions, and away from the use of general rules as decisional criteria. This is not surprising; it may make great sense as a policy matter to apply short-cut presumptions, but in any given case it may well produce the wrong result on those particular facts. Judges are very sensitive to injustice and are in general likely to be more concerned about the litigants before them than about the broad policy implications of their aggregated individual decisions. In other words, they like to get it right, and "it" is most often the specific matter in front of them.

The 1982 Merger Guidelines revisions began to reflect this movement toward more careful analysis of the specific facts, and the 1992 and 1994 revisions to the Merger Guidelines essentially abandoned the Philadelphia National Bank presumption, treating concentration as merely a starting point for analysis. Those Guidelines adopted the "tell me a competitive story" approach, insisting that market structure is part but not all of the story, and demanding an explanation for why this transaction in this market at this time was likely to generate anticompetitive effects. Since the agency-sponsored Guidelines have such influence over courts and practitioners, the last decade has seen merger analysis move steadily in that direction. Indeed, the recent publication by the agencies of actual enforcement data shows that the result has been a much more nuanced approach to merger analysis, with many transactions that raised questions under a market share or structure screen eventually not challenged after more detailed factual analysis. And the government has also adjusted in the courtroom — to rely more on documents, customer testimony, and other factual material, not just recitations of market shares or broad economic principles (although these are frequently included and sometimes relied on heavily). This shift in approach is no doubt part of the reason for the smaller number of merger challenges over the last decade or so. It is inevitable that agencies are less likely to bring cases where the resolution depends on factual analysis, since fact-based prosecutions are much harder than those based on presumptions. It probably also explains some of what some critics of the agencies see as their less aggressive approach to merger enforcement over the last two administrations. This is why the most successful advocates before the agencies in recent years have concentrated on attacking the chances that the agency could prevail in court, as opposed to arguing that the transaction was necessarily benign (approaches that are not necessarily inconsistent, but require somewhat different advocacy).

It is inevitable that this change in approach would make it harder for the agencies to win litigated merger cases. Thus, it is particularly impressive that, until very recently, the record of the federal agencies in litigated merger cases has been quite good over the last few years. Again ignoring hospital cases, the agencies have won far more than they have lost — until the last two months. Over that period, there has been a string of three straight losses — one by the FTC and two by the Antitrust Division — that may signal a real inflection point in merger enforcement. These three opinions all, in slightly different ways, give very little weight to presumptions and make it clear why actually having to prove that a merger will likely produce anticompetitive effects is hard — especially if the government’s burden of proof is seriously applied and it does not have the help provided by presumptions. Of course, it is possible that these decisions simply represent bad judge draws for the agencies, or perhaps weak prosecutions, or maybe the decision to litigate was just not supported by the facts. But it is at least worth considering the possibility that these three losses in quick order, from very diverse courts, may signal a new level of resistance by district judges to some of what have been the ordinary tools of federal merger litigation.

The most visible of these cases is US v. Oracle, decided earlier this month in the Northern District of California. There, the Antitrust Division was trying to block the acquisition of PeopleSoft by Oracle, arguing that the two were each other’s closest competitors in particular kinds of enterprise software. On the surface, this looked like a fairly straightforward (indeed easy) case for the Antitrust Division — the two companies were vigorous competitors, and it was generally accepted that there were few (maybe no) others perceived by at least some significant number of customers as real alternatives. In fact, the Division had lined up a series of sophisticated customers to testify that the products from these two companies were the only products they had or would consider for the particular functions involved, and that a 10 percent price increase would not change that conclusion. While there were others offering various pieces of the software suites offered by Oracle and PeopleSoft, the common perception was that at most one other company (SAP) offered truly competitive software suites in the functions at issue, and that SAP was a pretty distant third in the American market. In short, this seemed to be a relatively classic unilateral effects case, where the competitive concern is that the combined company will be able to raise its price unilaterally after the transaction because a significant number of customers do not see any acceptable options.

The 164-page opinion recently issued by Judge Walker rejects every important element of the Antitrust Division’s case. What may be more significant in the long run is how he did it. Judge Walker has considerable antitrust experience, including serving for some time as the Judicial Liaison to the ABA’s Section of Antitrust Law. In addition, he is technologically proficient, having been an early adopter of various electronic courtroom techniques. He had no trouble understanding the software concepts involved, and his lengthy opinion shows a detailed familiarity with the antitrust concepts at issue. Since the Antitrust Division has now decided not to appeal, this opinion will be cited by merger defendants for a long time. In Oracle, the Court accepted the concept of the unilateral Oracle effects theory of liability but severely constrained its use, concluding that it was only applicable where the parties involved were so distinguished from other competitors that they were, in effect, in a separate market. Perhaps more importantly, the Court completely rejected the customer testimony proffered by the Division, asserting that it simply reflected preferences, not careful analysis of what was possible if the acquisition was allowed to proceed. (Whether this is an accurate description of the record is open to serious question, but since there will be no appeal, this will be the conclusion that stands.) The Court also rejected the testimony of the government’s expert witnesses as based more on subjective conclusions than quantitative analysis. Since the government’s experts included one of the most experienced and respected antitrust economists, Ken Elzinga, who is also an experienced witness, it seems hard to imagine that they would not know how to go about economic analysis of the transaction. Judge Walker, however, found their testimony unpersuasive. And finally, the Court emphatically held the Division to its burden of proof, refusing to allow any presumptions to provide a helping hand.

Oracle follows by about one month the loss by the FTC (and several state Attorneys General) of their bid, in FTC v. Arch Coal, Inc, to gain a preliminary injunction in the District Court of the District of Columbia against the combination of two coal producers. Arch Coal had sought to acquire Triton Coal. Both operate mines in what is known as the Southern Powder River Basin region of Wyoming. All parties agreed that SPRB coal was a relevant market, but the FTC also argued that within that market, higher BTU content coal was a separate and distinct market within which to evaluate the combination. This was significant because only four companies (including the merging parties) produce high BTU coal from the SPRB, while there are seven producers of SPRB coal in total. Judge Bates rejected the FTC’s market definition argument, in large part because (at least in the Court’s view) even the government’s expert was less than convinced (and thus less than convincing) on this point. So far, nothing extraordinary (save perhaps the preparation or choice of the witness). But the Court went on to find that the level of concentration in the SPRB market was sufficient — barely, but sufficient — to support a prima facie case. For some courts in the past, this in itself would have been enough, under Section 13(b), to entitle the FTC to a preliminary injunction. But here, the Court went on to consider the parties’ rebuttal to this prima facie showing, and in so doing, very strictly held the FTC to its burden of proof, ultimately concluding it had not been met.

The FTC alleged, and the Court found, that the SPRB market was potentially susceptible to coordinated interaction, and the Court also agreed that there had been attempts by Arch Coal in the past that might be construed to encourage just that kind of activity. In such a context, the FTC argued, it would be profitable for the remaining producers to coordinate to reduce output, and the fact that at least some of them had already tried in the past made it logical to assume that they would try again in the future. While one might think that under these circumstances meeting the 13(b) standard would be relatively easy, the Court described the FTC’s theory as "novel" because it was focused on output restriction rather than price coordination, and it ultimately found that the FTC had not met its burden of proof. What is striking about Arch Coal is that the Court basically accepted the FTC arguments, but nevertheless concluded that they were not sufficient to meet its burden under 13(b). And like in Oracle, the Court Oracle refused to rely at all on testimony from utility customers concerned about the transaction, concluding that "[c]ustomers do not, of course, have the expertise to state what will happen in the SPRB market . . . ."

The complete rejection in both Arch Coal and Oracle of what has become a standard evidentiary tool for enforcers — customer (or putative "victim") testimony about how the transaction is likely to harm them — is a real blow to both agencies’ current approach to merger litigation. The views of customers have carried great weight in the agencies’ decision- making processes and have been a critical part of most successful prosecutions; indeed, when the agencies have been unable to muster witness testimony against the transaction, they have had great difficulty in winning in court. Now, we have two cases where the agencies were not only unable to win despite strong customer testimony, but where the courts essentially disregarded it, in almost the same way they might discount competitor views. This is bizarre, since the last time we looked, the antitrust laws were intended to protect consumers, and the fact that consumers (or customers in these cases) might have their own self-interest in mind when expounding their view about the transaction is exactly the point. It might well be reasonable to give little weight to customer testimony that seems to have no basis and is inexplicably contrary to the other evidence in the case, and perhaps an argument can be made that that is what happened here (in Arch, at least). But in both cases, it appears on the surface that the court allowed its view of economic logic to completely override the perspective of the people that participate directly in those markets and that would have to respond to any changes in market structure and performance. If these decisions mean that customer testimony no longer has the evidentiary significance it has had in the past, this will have significant implications for merger enforcement.

There is another piece of bad news for the agencies in Arch Coal. After the FTC began its investigation of the trans- Coal action, Arch Coal agreed to sell one of the two mines it had agreed to acquire from Triton to a third party, Peter Kiewit Sons, Inc. The FTC took the position that it usually takes — it was entitled to evaluate the original transaction, not the amended transaction. The Court rejected this position and agreed to evaluate the transaction before it, including the subsequent sale to Kiewit. From the FTC’s perspective this is anathema, because it allows the parties to react to an FTC challenge by changing the deal in mid-stream. In addition to the practical difficulties this presents to the FTC staff — having to switch gears in the middle of an investigation — the FTC also dislikes the notion that the parties can decide how to adjust a deal to respond to antitrust concerns. Like any bureaucratic institution, it would prefer to control the remedy to an antitrust concern, not leave that in the hands of the transgressors. Many practitioners have long believed that the FTC’s position on this issue was not sustainable, and this decision and that in DFA discussed below confirm that view. In Arch Coal, the acquiring company adjusted the deal Coal after the FTC began its investigation and the Court evaluated the adjusted deal, not the original transaction. In that case, this meant that the number of independent producers in the market found by the Court was not reduced at all, which was obviously not helpful to the FTC’s case.

The Antitrust Division found itself on the same end of the stick in US v. Dairy Farmers of America, in the Eastern District of Kentucky. There, DFA (a milk marketing cooperative) formed an LLC with some local interests and acquired a fluid milk processing plant in Somerset, Kentucky. DFA owned a 50 percent interest in the LLC. DFA already owned a 50 percent interest in a partnership that owned a milk processing plant in London, Kentucky, a competing processor. The Antitrust Division sued, alleging that the common ownership by DFA would reduce competition between the two processors. Following the suit, both the LLC and the partnership were restructured, with DFA’s interest in the LLC becoming non-voting, and the management structure in both entities was clarified to establish that DFA had no operating role in either. DFA then moved for summary judgment, arguing that on the facts of the restructured arrangements, its non-voting interest in the LLC did not violate Section 7.

The Antitrust Division objected to evaluating the adjusted deals, but like the FTC in Arch Coal, it lost. It then argued Coal that, notwithstanding all these maneuvers, the fact remained that DFA now had a 50 percent ownership interest in two significant competitors, and it could be reasonably assumed that this common ownership would affect the intensity of the competition between the two processors. There have been times in the past when this would have been an easy win for the Division. Indeed, smaller (sometimes much smaller) common stock holdings have been asserted (by both the Division and the FTC) to raise competitive concerns, even when the holdings were non-voting; Time Warner - Turner and Northwest - Continental come to mind. In addition, of course, this was hardly the fi x sought by the Division, which wanted to eliminate DFA’s common ownership, not simply adjust it. But Judge Forester was not sympathetic.

The Court first concluded that the Philadelphia National Bank presumption did not apply because DFA did not control both competitors. Thus, the Court held, the Division must come forth with real evidence that would defeat the summary judgment motion. The Division responded with economic evidence, arguing that since DFA was entitled to 50 percent of the profits of each, it had a particular incentive and opportunity to influence or pressure each competitor to reduce its competition and thus increase its profits, and therefore that the transaction "makes it easier to lessen competition between the two dairies, either through tacit means or otherwise." The Court rejected this argument, basically relying on the parties’ representations about how competitive decisions were actually made by the parties and the absence of any voting rights by DFA in the LLC. "Every investor, however small, has an incentive to achieve higher profits and perhaps even to communicate with management on these issues," said the Court, "[b]ut this obvious point does not establish the probability of anticompetitive effects that would render the investment illegal under Section 7." So much for economic theory, standing alone.

These three opinions have plenty of arguable flaws, and it remains to be seen whether they will withstand the passage of time. But they do contain some common threads that should raise real red flags for the enforcement agencies and perhaps offer some additional reason for the Antitrust Modernization Commission to look at the system of US merger regulation. First, they give only lip service at best to the Philadelphia National Bank presumption that increased concentration implies less competition. While this has been the cornerstone of modern merger analysis, the agencies long ago came to this same place in their internal analyses, but they have been reluctant to abandon the litigation advantage of citing PNB and other decisions from three and four decades ago, decided on the basis of completely different analyses than apply today. Perhaps these opinions will finally convince the agencies that it is time to upgrade the litigation complaints and briefs to reflect more of the actual internal analysis, since it does not look like the old techniques will be successful.

Second, and potentially of even greater significance, both Oracle and Arch Coal essentially ignore customer testimony, which has become a critical part of the evidence (along with bad documents) in most government merger cases. If customer testimony is to be trumped by more "objective" analysis, this will require a major shift in the way the agencies prepare cases. Today, they basically depend on documents backed up by customer testimony and economic evidence. Documents are still likely to be highly probative, but properly counseled companies will produce fewer useful documents. If customer testimony is going to be viewed with skepticism, and economic evidence (even from the most respected economists) treated similarly, the road to winning merger cases will get much harder for the agencies. As an aside, it is interesting to contrast this state of affairs with the EC, where great weight is given to competitors, even if (and sometimes especially if) the transaction will produce significant efficiencies that would presumably put great pressure on competitors — to the benefit of consumers. The European Commission appears to be willing to sacrifice what it views as short-term consumer benefits to protect against long-term "competitive dominance." These decisions generally reject concerns about both long- and shortterm competitive effects even where consumers feel strongly enough about the transaction to stand up and oppose it.

And finally, in both Arch Coal and DFA, the Court allowed the parties to adjust the deal and performed its analysis on the adjusted deal — in one case after the investigation began and in the other, after the case was brought. If this approach holds, merger parties have a lot more tactical options than they have had in the past.

Are these cases just coincidental outliers, or do they tell us something about the likely attitudes of many judges to antitrust merger challenges? Northern California, the District of Columbia, and Kentucky — these are pretty diverse venues, and the similarity of the judicial approach in these very different jurisdictions should be of concern to the agencies. If these are not outliers but leading indicators, what does this mean for future merger enforcement? Interesting questions to ponder.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.