Some proponents of more vigorous antitrust enforcement have expressed concern that these losses show it is too hard to defeat potentially anticompetitive mergers. One view is that proving a Clayton Act violation has come to demand too much economics—a kind of “CSI effect,” in which economists are held to too high a standard. Another view is that courts pay too little attention to the economics. This view sees courts as de-emphasizing the more general economic teachings underlying Section 7 analysis, especially the role of consolidation in changing the incentives of the merging firms, and instead as being swayed too easily by the potentially self-serving testimony of company business executives about how they plan to behave post-consummation. These viewpoints, which are not inconsistent and are sometimes expressed simultaneously, contend that judges ask for too much from economics while at the same time giving the broader insights that economics can provide too little weight.
To overcome these concerns, some have suggested lowering the economic bar that plaintiffs must clear. They argue that a sufficient amount of economics is only that necessary to explain at a high level—along the lines set forth in the Horizontal Merger Guidelines —why antitrust has traditionally been concerned with the changes in incentives arising from horizontal and vertical consolidation, especially when the transaction eliminates a maverick or strengthens a leading industry firm. The extreme version of this position would put the burden of proof on the merging parties, with all horizontal mergers satisfying certain criteria deemed unlawful unless the merging parties could prove that they would not lessen competition.
We examine these cases from a slightly different perspective. Instead of considering whether they were rightly decided, or whether or how the law should change to compel (or allow) different outcomes, we ask what these cases tell us about what makes economic testimony compelling within the current legal framework. There is no doubt that government merger cases —pled on credible markets and supported by effective economic presentations that are woven closely with other evidence—can, and routinely do, prevail even in the face of challenging facts. One recent example is Peabody/Arch Coal, where the FTC prevailed against coal mining companies despite sharp declines in coal demand precipitated by customers (electricity generators) switching away from coal to natural gas and renewables. The manner in which courts assessed the economic consequences in the recent losses, and were left unpersuaded by the economic testimony put forward by the government, has lessons for the future role of economics in complex merger litigation.
The Rise of Economics in Enforcement and Antitrust Litigation
Before we discuss each of the recent losses, we review how economic testimony came to its present role in merger litigation. Given the Clayton Act’s requirement that the government prove a substantial lessening of competition, it is not surprising that economics plays an important role. Testimony about economic effects took on added importance when the Supreme Court moved away from a focus on static industry structure in General Dynamics.
In the 1980s, economics took deeper root at the agencies, both in deciding whether to bring cases and eventually in litigating those cases in court. The first Merger Guidelines, introduced in 1968, were not particularly heavy on economics. They emphasized market structure and included structural presumptions much stricter than today’s standards. The Guidelines were first revised in 1982, introducing the hypothetical monopolist test and the Herfindahl-Hirschman Index. Further revisions continued to reflect the agencies’ evolving experience and practice, correspondingly expanding the scope of the inquiry into the more nuanced analysis that we recognize today—one focused much more on the ultimate economic effects of a transaction than the static effect it will have on market concentration. For example, the Guidelines evolved to distinguish more clearly between coordinated and unilateral effects and to give a more explicit role to evaluation of entry and efficiencies.
The agencies’ practice also evolved over time, leading to the 2010 Guidelines, still in effect, which emphasize a merger analysis toolkit that is even more heavily skewed towards economics. For example, the 2010 Guidelines stress that market concentration is only the starting point of the analysis (while raising presumption thresholds), note that some of the analytical tools used by the agencies do not rely on market definition, and explicitly refer to merger simulation and upward pricing pressure analyses as among the tools that economists will bring to bear in evaluating proposed mergers.
While some might have been concerned that the increased emphasis on economics would make enforcement more difficult, the opposite seems to be true. Until the recent set of losses, the government consistently won cases litigated after the 2010 Guidelines were issued, with economics playing a key role and the Guidelines often cited in support of the decisions. Of course, other factors were at play during this time and it is not clear that the 2010 Guidelines were the only or main driving force behind these wins. Nevertheless, the 2010 Guidelines seem to have helped courts overcome points of confusion that had contributed to the rejection of economic testimony and government losses in the late 2000s by clarifying the role played by different types of evidence and the relationship between market definition and competitive effects.
What Really Happened in Recent Government Losses?
This brings us to the recent government losses. It is tempting to look at these cases and draw broad conclusions about the declining impact of government economic testimony. However, a detailed examination of the cases leaves the picture far less clear. Each of the cases presented its own unique set of facts, and plaintiffs marshalling those facts pushed enforcement boundaries in different ways. Accordingly, though the government-side economic showings were criticized or minimized along the way to unfavorable decisions, the specific reasons were different.
AT&T/Time Warner
AT&T/DirecTV, the nation’s largest distributor of traditional subscription television content attempted to acquire Time Warner, which was a major content provider (or programmer) through its ownership of many of the country’s top TV networks. In November of 2017, the DOJ filed a complaint to block the acquisition. The DOJ’s main concern was that a vertically integrated programmer would have greater incentive and ability to threaten to withhold programming from rival distributors so as to demand higher prices and more favorable terms. The specific economic thesis behind DOJ’s challenge was that post-merger in the event negotiations between Time Warner and rival distributors reached impasse, some of the rival distributor’s customers might divert to AT&T/DirecTV, thereby lessening the costs of reaching impasse and increasing the leverage of the vertically integrated firm. At a high level, this theory posed two challenges for the government’s expert economist: getting the court’s “buy in” to the (somewhat) novel theory and quantitatively showing that its likely effect would be “substantial.”
To advance these points, the government’s economics expert presented a bargaining model that aimed to show the increased incentive (and leverage) AT&T would have post-merger in negotiating with cable systems for Time Warner programming. The district court criticized both the bargaining model and the specific findings of the model and ruled for the defendants. The court rejected the theoretical model, referring to it as a “Rube Goldberg contraption.” The court also rejected the quantitative predictions from the model, finding them to be rebutted by executive testimony and economic evidence. In particular, the government’s economist had left unrebutted testimony by one expert that undercut certain inputs used by his model and a regression analysis presented by another expert to support testimony that the combination of Comcast and NBC Universal in 2011, which also combined a distributor (Comcast) with an owner of video programming (NBCU), did not have adverse price effects. The Comcast/NBC transaction had been remedied by the DOJ via a consent decree that imposed a mandatory arbitration obligation that the defendants in AT&T/Time Warner argued was analogous to the commitments they had already made to programmers.
The DOJ appealed the district court’s opinion, arguing among other things that it ignored or misapplied economics. The D.C. Circuit affirmed, but in the process did not reject the validity of the theoretical model put forward by the government’s economist. Instead, it observed that the district court had accepted the economic theories underlying the merger simulation but rejected the simulation’s predictions as applied to this particular case, in large part due to the defendants’ econometric rebuttal. In that sense, the court here did not ignore economics, but merely chose to credit the defendants’ economics over the plaintiff’s and to reject the plaintiff’s view of the merged firm’s incentives based on economic and other evidence.
Evonik/PeroxyChem
The proposed merger was a combination of manufacturers of hydrogen peroxide, a chemical that comes in various forms for different uses. The crude form of hydrogen peroxide is purified into standard, specialty, pre-electronics, and electronics purity grades. The FTC’s theory was a fairly “standard” one that the transaction would increase concentration in an allegedly concentrated nationwide market for all non-electronics hydrogen peroxide, thereby yielding unilateral and coordinated effects that would allegedly lessen competition. The difficulty for the FTC was that the merging parties largely operated in different regions and different product segments. If these separate segments were the proper antitrust markets, then there was little to no overlap.
To deal with this difficulty, the FTC defined a market that included all non-electronics grades and, in a novel move, argued that supply substitution broadened market boundaries despite the lack of functional interchangeability from a demand-side perspective. Specifically, the FTC included all non-electronics grades in the market not because customers actively substitute between standard, specialty, and electronics grades, but because hydrogen peroxide producers could “swing” production across the various grades. This approach deviated from the general principle that market definition focuses solely on demand substitution. A footnote in the Guidelines does allow for supply substitution to influence market definition, but only under specific conditions and only as a matter of analytical convenience. The court ultimately found that the facts of the case (including the economic testimony) did not meet the specific conditions required by the Guidelines for the FTC’s proposed exception to demand-based market definition.
While the court rejected the FTC’s market definition, it did not reject its economic analysis. It relied on testimony from both the plaintiff’s and the defendants’ economic experts to evaluate whether market realities fit the conditions set out in the Guidelines for supply substitution, but ultimately sided with the defendants’ expert.
T-Mobile/Sprint
This was a horizontal merger between two of the four major wireless carriers in the United States. As a result of conditions imposed by the FCC and agreed to in a consent decree with DOJ, the transaction also promised a fourth nationwide player through the divestiture of certain assets to Dish. The litigation was notable for the detailed efficiencies evidence that defendants brought to bear and, by contrast, the relatively limited economic evidence presented by the plaintiffs. Despite ruling for the plaintiffs on market definition (relying on the plaintiffs’ economic expert), the court nonetheless found that the presumption of harm arising from the merger’s impact on concentration—and the characterization of the deal as a “4 to 3” merger—was insufficient to carry the plaintiffs’ burden and that the plaintiffs’ other evidence of competitive effects was likewise insufficient.
Especially considering the lengthy pre-litigation investigation to which this merger was subjected, it is noteworthy how heavily the plaintiffs’ case relied on the Philadelphia National Bank presumption rather than any form of quantitative economic analysis. Beyond market shares and concentration measures, the plaintiffs’ economic expert primarily just explained and applied the Guidelines framework for considering coordinated and unilateral effects. The coordinated effects evidence amounted to little more than a checklist of factors (disputed by defendants’ expert) and an illustrative calculation of harm. The unilateral effects evidence focused on an upward pricing pressure calculation that appears to have struck the court as divorced from market realities, especially its failure to account for the impact on post-merger incentives of the large expansion in capacity the transaction would achieve.
At trial, the defendants chose to make much of their rebuttal case through executives. The trial featured extensive executive testimony on the merger rationale, engineering testimony on network planning and projected efficiencies, and testimony from Dish executives about their plans to enter the market. This was not expert economic testimony, but it was focused squarely on economic incentives: the inevitable network combination would reduce costs, expand capacity, and improve quality, and the merged firm would have the competitive imperative to use these new tools to gain share at the expense of the two leading firms, Verizon and AT&T.
This did not mean that economists played no role in the defendants’ litigation strategy and success. The defendants still needed an economist to translate the parties’ engineering testimony into marginal cost reductions and to put a monetary value on the anticipated network quality improvements. Even if there was no sophisticated plaintiff analysis to rebut, the defendants also needed an economist to rebut the plaintiffs’ arguments on concentration and competitive effects.
Overall, T-Mobile/Sprint appears to be a case where the court was persuaded by the defendants’ evidence on the nature of wireless competition, anticipated efficiencies, the reduced strength of Sprint going forward, and the likely strength of the entrant and divestiture buyer. Economic testimony was insufficient to sway the court to the plaintiffs’ view of the merged firm’s economic incentives, yet this was at least in part due to the economic testimony offered by defendants both at and before the trial.
Sabre/Farelogix
The acquirer in this case was Sabre, a global distribution system (GDS) that operated a large platform for booking, ticketing, and other transactions between airlines and travel agents. The target was Farelogix, a smaller firm that provided technology services to airlines. The DOJ alleged that Farelogix’s services were an important means for airlines to facilitate transactions with travel agents, whether by bypassing the GDS or by disintermediating it (i.e., substituting some of the services GDSs offer with Farelogix’s technology). DOJ referred to the relevant services that Sabre and Farelogix offered—whether as a standalone service or as part of a bundle—as “booking services.” The DOJ claimed that the elimination of Farelogix as a competitor for these services may lead to less innovation, higher prices, and a reduction in the bargaining leverage for the airlines when negotiating with the GDSs.
The case was a difficult one for the DOJ because it challenged an acquisition by a large, established technology firm of a company that had heretofore played a specific and small competitive role, but had the potential to become a stronger competitor in the future. Despite evidence of past competition between Sabre and Farelogix, establishing a structural presumption was challenging because of Farelogix’s small size. Furthermore, the bundled product offered by Sabre—with Sabre not offering a separate “booking services” product and consequently there being no well-defined price for that set of Sabre’s services—made it difficult for the DOJ to make certain showings that are customary in merger cases, including computing a SSNIP on Sabre’s service as part of the Merger Guidelines’ market definition exercise.
The court ruled for the defendants and was critical of the plaintiff expert’s economic testimony. The court seemingly struggled to reconcile the evidence it was shown. On the one hand, it found a clear factual record of competition between Sabre and Farelogix, evidence that Sabre will have “the incentive to raise prices, reduce availability of FLX OC, and stifle innovation,” and several defense witnesses showing a “surprising lack of credibility” on multiple points. On the other hand, the court noted a lack of concern about the merger among many industry participants and found unconvincing the alleged “booking services” product market definition. And, viewing Sabre’s GDS as a two-sided platform, it felt constrained by the Supreme Court’s holding in Amex as compelling the conclusion that Farelogix, as a one-sided firm, could not compete with Sabre as a matter of law.
On its face, the opinion recognized the tension in its findings but left unresolved numerous apparent contradictions. Ultimately, the court concluded both that the “evidence suggests that Sabre will have the incentive to raise prices, reduce availability of FLX OC, and stifle innovation,” and also—in the very next sentence—that the “DOJ has not persuaded the Court that Sabre will likely act consistent with its history or these incentives and actually harm competition if it is permitted to complete the acquisition of Farelogix.” The court was aware that the ruling for the defendants would “strike some . . . as odd” since “[o]n several points that received a great deal of attention at trial . . . the Court is more persuaded by DOJ than by Defendants.” But the court repeatedly emphasized that it was the DOJ’s burden to prove its case, and observed that the DOJ and its economic expert failed to do so persuasively.
Jefferson/Einstein
This case involved a merger of hospitals in Northern Philadelphia and surrounding suburbs. Unlike many of the hospital mergers the FTC has successfully challenged in recent years, this transaction involved hospitals at the edge of a large and densely populated metropolitan area. There were many health care options in the southeastern Pennsylvania region surrounding Philadelphia, including more than ten hospital systems and dozens of hospitals dedicated to general acute care, several of which were well-known academic centers. The FTC’s recent hospital merger playbook, moreover, has featured harm to commercial health insurance providers arising from hospital consolidation, yet in this case those insurers were themselves highly concentrated, with the leading insurer—a key FTC witness—enjoying over 50 percent market share and having a history as a tough negotiator with providers.
The court ruled for the defendants on both market definition and competitive effects, rejecting both the economic and insurer testimony offered by the government. The court was persuaded by the evidence the defendants presented on the robust hospital competition in northern Philadelphia and the many options that were available to patients (and thus their insurers) just beyond the alleged market boundaries. The economic testimony offered by the government was insufficient to overcome these difficult facts and, perhaps not surprisingly, ended up being criticized (along with insurer testimony) as inconsistent with commercial realities.
Nonetheless, in rejecting the economic testimony, the court also appears to have rejected economic concepts that have been accepted by other courts. For example, the court appeared unpersuaded that the government could rely on patient preferences (in the form of patient diversion ratios) to infer insurer preferences. The government economist’s contention that such a link was “pretty basic economics” did not move the court, nor did his testimony that most economists would agree that when costs go up in imperfectly competitive markets there is some pass through. The court brushed aside this testimony as “academic econometric analysis” that was unsupported “by credible evidence.”
Even if these concepts were indeed basic and generally accepted, it appears that the court was not persuaded by assertion alone. In particular, the court was aware that more direct evidence linking patient preferences to insurer demand had been offered in other hospital merger cases, but—absent similar evidence about the markets at issue in this case—the court was not willing to view that evidence as sufficient to establish a general principle that insurer preferences depend on patient preferences. While some of the evidence the court appears to have wanted to see may simply not have been available (for example, an appropriate natural experiment), more analysis and, at a minimum, taking more time to teach bedrock economic concepts might have been helpful to the government’s case.
These Decisions Do Not Portend a Diminished Role for Economics
At least four themes emerge from the courts’ treatment of economics in these cases. First and foremost, in each case the central factual question asked by the court was an inherently economic one: was the likely effect of the transaction anticompetitive? Economics thus remained central to the inquiry, even though the courts were not persuaded by the conclusions reached by the government experts. This dynamic is unlikely to change unless the law moves towards a more structural standard that entitles the government to a stronger or more readily available presumption of harm.
Second, there is nothing surprising about federal courts weighing different types of evidence and insisting that the economic conclusions drawn from the facts be consistent with other evidence and market realities. Despite the importance of economics to the application of the antitrust laws, economic experts have no special role in antitrust trials (much to the disappointment of economists). Like any expert witness in a federal case, an economist’s role is to assist the trier of fact in his or her job of applying the law to the facts of the case, and when economic conclusions seem at odds with key facts—as in AT&T/Time Warner and Evonik/PeroxyChem—the facts will trump the economics. This conclusion is reinforced by the fact that many other courts have relied on complicated economics presented by antitrust plaintiffs. For example, the court in Aetna/Humana likely was willing to rely on complicated merger simulation methods in part because the plaintiffs had ensured that this evidence was developed and presented in a manner fully consistent with the non-economic evidence and testimony at trial.
Third, given that merger litigation typically requires predictive judgments about likely future conditions, some factual contexts will be more difficult than others for lawyers and economists alike. Courts are rightfully suspect of “crystal ball gazing” and prefer to make judgments based on forecasts firmly rooted in observable experience. Economists testifying for government plaintiffs naturally face somewhat greater challenges persuading courts of their predicted effects when the target had a limited track record or markets are changing fast (as in Sabre/Farelogix), or where past experience is not representative of the market conditions before the court (as in Jefferson/Einstein).
Fourth, the challenge of making economic predictions against the backdrop of a complex factual record is not insurmountable, provided the economic testimony accounts for the challenges. We cannot predict whether any of the five outcomes might have been different had the economics been presented in another way, but several practical lessons can be learned from what did not work in those cases.
Help the Court Interpret the Significance of Other Evidence
The recent government losses highlight the vital importance of acknowledging and addressing evidence in the case offered to refute the plaintiffs’ theory of harm. Over and over in these cases the force of the economics was diminished by perceived failures to contend with unhelpful facts: the parties’ commitment to arbitrate content disputes that reached impasse in AT&T/Time Warner, the compelling impact on pricing incentives of the network capacity expansion forecast in T-Mobile/Sprint, the disincentives of producers to swing their output between grades of hydrogen peroxide in Evonik/PeroxyChem, and many more examples. Economic theorizing will be rejected as unpersuasive—or simply ignored—when the court perceives it as out of step with other testimony or inconsistent with market realities. As a corollary, no amount of economic modeling is likely to save a weak case.
Put Forward Compelling Quantitative Analyses. Courts will be looking for reasons to trust the predictions offered by the government’s economists, especially where there is a strong economist on the other side of the case and the factual record is complex or conflicting. Well-credentialed economists do not automatically cancel one another out, as perhaps best shown by the weight placed by the court in AT&T/Time Warner on regressions offered by one former DOJ chief economist testifying for defendants against modeling offered by another. To be persuasive, economic testimony against a merger should offer more than conclusory restatements of basic principles, simplistic calculations of market shares or GUPPIs, or checklist-style review of the factors bearing on the risk of coordination, as in T-Mobile/Sprint. Mining the documentary record—and thus the parties’ own words—has become a routine source for confirmation for predictions based in economics. But with documents pointing in many directions, and the authors of those documents potentially testifying at trial, an economic case rooted primarily in the documentary record may prove to be fragile.
Economists will be at their most persuasive when they are marshaling hard data that supports the economic inferences that they are drawing from the facts. This need not be in the form of complex econometrics—indeed, greater complexity can risk undermining the court’s appreciation of the simple insights. Natural experiments can be as straightforward as “look at what these companies did when confronted by similar market stimuli in the past.” Especially in data-rich environments, there may well be a need to go further—showing how pricing or other competitive outcomes are affected by differences (or changes) analogous to those flowing form the transaction.
The absence of such evidence at trial may carry with it an inference that it does not exist, as seemingly occurred in Jefferson/Einstein and perhaps in T-Mobile/Sprint. We are living in an increasingly data rich environment and it would not be surprising were courts to expect that data be used in building an economic argument. The flip side is that credible data analysis can be expensive and potentially puts budget-strapped agencies at a disadvantage. Furthermore, courts need to recognize that even in a data rich environment, a credible empirical analysis is not always possible.
Take the Time To Teach
There will inevitably be cases where the competitive predictions are not straightforward and quantitative analyses are not readily available. But the economist’s toolkit is not empty in these situations. They should take the opportunity to teach. Some of the government losses show the perils of treating the economist presentation as an exercise in putting economic conclusions into evidence in the shortest time possible. Those presentations can be so much more powerful if designed from the outset to go beyond “check-the-box” case building. Much like a persuasive businessperson can educate the court about the realities facing market participants and breathe life into a procompetitive business plan, an economist in a merger trial can bring a sharp-edged perspective to the factual record.
Teaching does not solely mean walking the court though basic economic principles and why they are important. It also means applying those principles to the myriad facts the court will have been immersed in at trial and guiding the court gently through all of them—good and bad—towards a predictive judgment about competitive effects that flows naturally from observed experience. Finally, educators know from experience that teaching is more effective when it is a conversation. Finding ways to make direct examination and re-direct more interactive, rather than just mechanically checking off questions, can help drive the insights home. Finding ways to engage the court so as to directly address any outstanding questions can make all the difference.
