Skip to main content

Deputy Assistant Attorney General Jon Sallet of the Antitrust Division Delivers Remarks at the American Bar Association Fall Forum


Washington, DC
United States

The Interesting Case of the Vertical Merger

Thank you.  And thanks as well to Barry Nigro, Kathleen Foote, Bill MacLeod and the American Bar Association Antitrust Section, for the opportunity to speak today at the Fall Forum.  This all came about over the summer, when I thought that it might be interesting to do a presentation on how the Antitrust Division has looked recently at vertical mergers.  Not a broadly engrossing topic, I thought at the time, but one that might be of some limited utility to antitrust practitioners.  But now there’s actually a vertical merger in the news.  To be very serious, the timing of this was coincidental, and this speech does not address any pending transaction in any way.  Nor do I want you to read into it any kind of coded messages.  It’s just an attempt to summarize recent developments in what Sherlock Holmes, in a different context, said was a question of “some interest.”  

My purpose today is to provide my personal views of the import of the division’s recent approach to vertical mergers and other mergers that raise the potential for vertical restraints on competition.  Of course, you may be familiar with the so-called Non-Horizontal Guidelines, which were issued in 1984.  But it is widely recognized that the competitive effects theories now applied by the division in assessing vertical and other non-horizontal mergers go beyond those articulated in 1984 and reflect more recent economic literature and practical experience on whether and how a vertically integrated firm would act to harm competition.  In other words, the division’s concern with possible foreclosure, raising rivals’ costs and other mechanisms for harming competition that can arise from such deals is substantially broader than what the 1984 Guidelines express.  Moreover, efficiencies are not always cognizable and remedies will not always be efficacious, issues the 1984 Guidelines do not adequately address.  

I recognize that some observers have suggested that the continued existence of the Non-Horizontal Guidelines means that the division does not devote many resources to the review of vertical transactions; I believe that this conclusion is belied by the recent work of the division, in completed reviews such as Comcast/NBCU, Google/ITA, UTC/Goodrich, Monsanto/Delta & Pine Land, the reviews of first Comcast’s and then Charter’s proposed acquisition of Time Warner Cable, and most recently, Lam/KLA.  The FTC has also challenged vertical mergers, for example in GE/Avio, Pepsi/PBG and Coca-Cola/CCE, under similar theories of harm to those I will be discussing today.  Indeed two decades ago then-Deputy Assistant Attorney General Steven Sunshine told the Spring Meeting that “vertical merger enforcement is an important part of the department’s merger policy,” and I think the record reflects that is still true today.    

The starting point of analysis should be this:  “vertical” is a term that describes the economic relationship between firms.  “Vertical” is a subset of the class of non-horizontal relationships where two companies operate at different levels of production or distribution—insofar as their relationship is vertical, they are typically not serving, or seeking, the same set of customers with the same types of products or services.  By “vertical,” I mean specifically the set of supply-chain relationships with a single firm present in both an “upstream” market of providers and a “downstream” market, usually of distributors.  (One recognizes that the use of the terms “upstream” and “downstream” rests on the perspective of the observer.  Distributors can be described as “downstream” of manufacturers, who create an input into the distribution, or as “upstream” of the manufacturing market, to which they provide an input.  For purposes of today’s discussion, I will use the former formulation and describe manufacturers as “upstream” of distributors).

The key point I wish to make is this:  “vertical” describes a business relationship; but the identification of that business relationship does not by itself render a judgment as to whether competition may be helped or harmed.  To answer that question, we must go beyond simply the vertical nature of the business relationship to pose a series of questions about markets and competitive effects. 

So let me address three issues: first, how the division has recently assessed vertical transactions considering potential competitive benefits and harms.  Second, how the division takes note of mergers in which an outcome may be to increase the risk of vertical restraints, for example through increased bargaining leverage.  Third, how we have recently been thinking about the use of conduct remedies in these kinds of transactions.  I’ll begin with some background applicable to all three.  


It’s been a long time since the Supreme Court last adjudicated a vertical merger.  That was in Ford-Autolite in 1972.  But in the late 1970’s the Supreme Court took a significant step in changing the previous view of vertical contractual relationships.  Although not a merger case, GTE Sylvania has long been recognized as very important in the analysis of vertical relationships, undoing the per se rule for nonprice vertical restraints and recognizing their potential to enhance competition.  But I want to suggest that GTE Sylvania and one of its successor cases, Leegin, stand for three principles that are reflected in the division’s recent work on vertical transactions. 

First, Justice Powell’s opinion in GTE Sylvania rested on the principle that facts matter.  Justice Powell, for whom I clerked, was very proud of this opinion; and, in emphasizing the importance of factual analysis, he was hewing close to a core aspect of how he saw the business of judging. From the division’s perspective, the result, of course, is a careful analysis of the specific circumstances presented.

Second, the court had a very specific reason for concluding that the facts in that case pointed towards a pro-competitive outcome.  As Justice Powell’s opinion explained, the vertical restrictions in that case “promote[d] interbrand competition by allowing the manufacturer to achieve certain efficiencies in the distribution of his products.”  So the fundamental point was not merely that the relationship was vertical, but that on the facts before the court the vertical relationship led to an outcome that increased competition in a market.  

Third, the 2007 Leegin opinion found similarly: vertical price restraints are no longer prohibited per se; their effects are evaluated under the rule of reason because they can benefit competition.   Here the analysis of the court reinforces and extends the understanding of the court in GTE Sylvania.  As with the Powell opinion, the Leegin court recognized that resale price maintenance could “increase interbrand competition by encouraging retailer services.”  But the court understood that, as always under a rule of reason, anti-competitive outcomes would also be possible, explaining that resale price maintenance, for example, could be used by a powerful retailer to forestall innovation in distribution that cuts costs or by a powerful manufacturer to curb the sale of competing products of “smaller rivals or new entrants.”  In these observations, we see the recognition of both input and customer foreclosure theories.

These three themes run through our recent analysis of vertical mergers.  We ask, again and again, what do the facts tell us about the potential impact of a new arrangement on what the GTE Sylvania and Leegin Courts called interbrand competition?  And we examine very carefully the potential for input and customer foreclosure.

In other words, I believe that, while we have sharpened some of our tools, the essential inquiry has not changed.  But what we are seeing may well have.  

It’s worth remembering the classic hornbook example of a vertical transaction.  A vegetable retailer buys a vegetable farm in order to be able to assure her retail customers that they will get uniformly high quality vegetables.  There are a lot of vegetable retailers and a lot of vegetable farmers (and low barriers to entry for both) and the hypothetical often assumes the acquisition improves the competitive strength of the company, offering consumer benefits that could not be achieved through contract.  The hypothetical may similarly assume benefits to the upstream vegetable grower, who gains a stable retail outlet around which to plan crop inventory and harvests.  With those assumptions, the transaction raises no antitrust concerns; in fact, it appears to be procompetitive. 
And here it’s worth emphasizing that vertical integration can create significant efficiencies that benefit suppliers, distributors, and consumers alike.  Antitrust experts of different stripes have recognized that vertical mergers can supply competitive benefits.  In Judge Bork’s famous “Antitrust Paradox,” he wrote that “vertical mergers are means of creating efficiency,” that “may cut sales and distribution costs, facilitate the flow of information…create economies of scale in management, and so on.”  Steve Salop, in two co-authored articles, has helpfully described a taxonomy of cognizable efficiency benefits, which include cost and quality efficiencies, increased investment incentives, circumstances in which a vertical merger might reduce the potential for coordination, improvements in design and production, and eliminating double mark up of costs.   

In many cases, the division has ultimately determined a vertical transaction would create efficiencies.  For example, in Google/ITA, Google suggested the transaction would give it a platform on which to develop new and innovative flight search services for consumers.  The division ultimately settled on a remedy that retained this benefit of the transaction.  Many of the vertical transactions cleared by the division have presented significant potential efficiencies that factored into our final decision.    

We have not always accepted claimed efficiencies, however. 

The reduction of double marginalization is a good example of the need for a careful scrutiny of claimed efficiencies.  As a matter of arithmetic, if two firms with vertically related or complementary products both have some market power, they may be able to lower the price charged to the downstream market by eliminating the above-market markup otherwise  charged on two separate products.  Vertical integration can solve this problem and benefit consumers.  But other factors may be important to consider.  For example the division’s competitive impact statement in Comcast/NBCU explains how, after reviewing documents, data, and testimony, the division concluded “much, if not all, of any potential double marginalization is reduced, if not completely eliminated, through the course of contract negotiations.”  Indeed, I think it is fair to say that an omni-present question in the recent completed reviews of vertical transactions is whether benefits are merger-specific or whether the same efficiencies can be gained through contracting.     

Of course the question, especially, when vertical mergers involve concentrated markets is the potential for harm to competition.  Let me turn to how that’s come up in some of our recent reviews.  

Competitive Effects 

The circumstances that have given rise to concerns in vertical merger reviews differ from the simple vegetable hypothetical.  As a starting point, we’ve seen concentrated markets, upstream, downstream, or both.  Sometimes we’ve seen upstream inputs of competitive significance, and even uniqueness, to other downstream firms.  Downstream opportunities may be foreclosed to upstream rivals.  In some cases we’ve seen the flow of competitively-sensitive information that tends to create unilateral or coordinated effects.  The hallmark of the inquiry, whatever circumstances we observe, is to look for power over a relevant market and examine how it may be enhanced or maintained as a result of the transaction.  It’s ultimately the question GTE Sylvania and Leegin addressed: does the merger threaten interbrand competition?  Like identifying efficiencies, assessing those threats often requires a fact-intensive inquiry.

Input Foreclosure and Raising Rivals’ Costs

Input and customer foreclosure theories arise from the fact that vertical transactions can create opportunities and incentives for firms to handicap rivals, and such actions can harm competition if they weaken the constraint that rivals impose on the merged firm’s market power (or, in some cases, the combined market power of a collection of firms that can coordinate on a higher price, lower output, or other non-competitive result).  The division’s UTC/Goodrich review in 2012 is a good example.  The transaction would have made UTC both a major producer of large aircraft turbine engines and the sole-source supplier of critical components to one of its leading engine competitors.  Our investigation revealed the merged firm would have had the ability and incentive to withhold or delay delivery of critical components, among other things, to that direct competitor.  An impact—here an adverse impact—on interbrand competition naturally follows from this kind of foreclosure—competitors without access to critical parts do not constrain market power as well as those who can timely and effectively bring competing products to market.  That problem was resolved through divestitures that also remedied more traditional horizontal concerns.

A similar concern arose in Comcast’s acquisition of NBCU, where Comcast was buying unique content that was an extremely valuable component of rival video distributors’ channel packages.  Comcast enjoyed market power in video distribution, and the investigation suggested it could weaken competitive threats by raising the costs of critical content to downstream rivals like competing video distributors.  Similar to completely foreclosing access to an input, raising its costs can decrease the ability of downstream competitors to constrain market power.    

Innovation Effects 

The concern in Comcast/NBCU extended not just to the current video distribution ecosystem, but to nascent online video rivals that were then beginning to disrupt and change the delivery model.  That added an important layer of analysis that sometimes arises in vertical transactions:  we look not only at existing products and distribution systems but at how innovation and disruption are changing them to consumers’ benefit.  The Comcast/NBCU decree not only sought to protect existing video rivals from foreclosure, but it was also designed to prevent the merged firm from foreclosing or raising the costs of developing business models with which online entrants would attack long-prevailing incumbent market power.  The prospect was that online distributors would enter and bring new forms of competition to established video-programming business models of the kind traditionally operated by cable companies.  We recognized in our Competitive Impact Statement that online entry was nascent but that the merged company might use its new-found assets to diminish its competitive significance.  

The division’s consent decree with Monsanto in its acquisition of Delta & Pine Land is another example of how innovation can factor into a vertical foreclosure analysis.  Monsanto developed genetic traits to put into its seeds, while Delta and Pine Land, also a seller of seeds, had a history of partnering with independent developers of traits, and was an especially important partner for those developers.  So Monsanto would be buying a company that was an important participant in the process of competing against Monsanto’s traits.  The division concluded that the merger would lessen competition in the development of cotton traits that would compete against Monsanto’s traits.  We ultimately entered into a consent decree with both divestiture and conduct remedies that reduced this risk while also preventing separate horizontal effects of that transaction.  You’ll see similar innovation concerns reflected in the competitive impact statement for Google/ITA, relating in that case to software platforms, and in our recent press release upon the abandonment of the proposed merger between Lam Research Corp. and KLA Tecnor. 

Competitively Sensitive Information Facilitating Coordination

In addition to potential foreclosure effects, we have also sometimes considered whether a vertical transaction will harm interbrand competition by facilitating coordination, such as through the exchange of competitively sensitive information.  We had an information concern in 2011 when GrafTech sought to acquire Seadrift Coke LP.  GrafTech is one of the largest producers of graphite electrodes in the world, and Seadrift Coke was the second largest supplier of a critical input—petroleum needle coke.  GrafTech already had a supply arrangement with an upstream rival to Seadrift, and based on a close examination of that arrangement and the companies’ businesses, we concluded confidential information would likely have flowed between competitors and facilitated coordination.  Our consent decree limited the flow of that information in several respects, in order to reduce that concern.  Our decree in the Google/ITA transaction had similar requirements, walling off aspects of Google’s business from the customer data available to ITA.
Other Theories of Harm 

There are other theories of harm in vertical transactions that are beyond the scope of these remarks, such evasion of regulation.  Steve Salop and Daniel Culley recently wrote a helpful article in the Journal of Antitrust Enforcement that sets out such theories, and Jonathan Baker has written about this as well in the context of the Comcast/NBCU deal, as have others.  These theories are worthy of examination and, in the right case, may be the basis for the division’s factual analysis.  

Vertical Mechanisms of Harm without Vertical Integration 

The second topic I’ll touch on briefly is the presence of vertical mechanisms of harm in mergers that don’t necessarily involve a combination of vertically-related assets.  That’s part of what we saw in Comcast/Time Warner Cable (TWC) and Charter/TWC. 

Both Comcast/TWC and Charter/TWC would have been mergers of geographically non-overlapping cable and internet networks, strictly speaking, and they were not predominantly mergers involving vertical integration of supplier and distributor like the vegetable grower hypothetical.  When Comcast/TWC was announced, commentators assumed that we would focus on geographic product markets for cable subscribers and the lack of overlap between the companies in those downstream markets.  We did look closely at those markets, but found the lack of geographic overlap was not determinative because the transaction increased the ability of the merged entity to take actions that harmed nationwide downstream rivals.  
In Comcast/TWC, the post-merger firm would have controlled nearly 60 percent of high-speed broadband internet connections nationwide.  Comcast would therefore have controlled a large proportion of the connections all internet content providers need to deliver content to household customers.  Comcast would have also had greater incentive and ability to harm rivals to its cable television business including online video distributors like Netflix or Amazon Prime, by, for example, charging even higher interconnection fees for access to customers or degrading the quality of service.  This concern about the cost and quality of upstream providers’ access to downstream customers arose even though Comcast merging with TWC did not primarily involve vertical integration.  Comcast ultimately abandoned the transaction after both the department and FCC expressed concerns along these lines.   

Charter’s acquisition of TWC raised similar concerns, although with smaller shares.  The consent decree that Charter agreed to last year illustrates the vertical nature of the theory of programming foreclosure in that case.  The decree prevents Charter from entering into vertical contracts with upstream programmers that would harm video rivals by limiting their access to programming that they would use to compete against incumbent pay television providers like Charter.  Whereas the Comcast/NBCU complaint focused on whether the merged firm would itself withhold content from rivals, the Charter/TWC complaint focuses on the ability of the merged firm to raise rivals’ costs through the use of bargaining power with independent programmers.  The decrees address a similar mechanism of harm notwithstanding the different structures of the transactions themselves.  


Let me turn to my third topic, the subject of remedies.  Where we have identified that a vertical transaction threatens interbrand competition, we must still consider how to resolve that concern, particularly where substantial efficiencies are also identified.  

In vertical transactions, observers sometimes assume that conduct remedies will always be available and sufficient.  But that is not the current practice of the division—if it ever was.  Indeed, while the Antitrust Division Policy Guide to Merger Remedies (2011) says explicitly that “conduct remedies can be an effective method for dealing with competition concerns raised by vertical mergers,” it adds that creating an appropriate remedy requires identification of the relevant competitive concerns and it warns that “[n]o matter what type of conduct remedy is considered, however, a remedy is not effective if it cannot be enforced.”  For example, the Policy Guide explains, “[r]emedial provisions that are too vague to be enforced, or that can easily be misconstrued or evaded, fall short of their intended purpose and may leave the competitive harm unchecked.”  Thus, the core question under the antitrust laws will always be whether conduct relief is adequate to eliminate the risk of anti-competitive harms.  To be employed, conduct remedies must be adequate to address identified risks, must be able to be monitored by the division or a court, and must be capable of being effectively enforced in a timely manner.  As the Policy Guide to Merger Remedies makes plain, the question in any case is whether such criteria can be met.  Some vertical transactions may present sufficiently serious risks of foreclosing rivals’ access to critical inputs or customers, or otherwise threaten competitive harm, that they require some form of structural relief or even require that the transaction be blocked.


We've talked about economics and evidence.  But we should end by acknowledging the importance of experience.  Courts call the Sherman Act a common law statute.  And a different Holmes, Oliver Wendell Jr., started his first lecture on the Common Law by reminding us that the evolution of law looks not only to logic but also to experience.  Today I have attempted to provide a review of the recent experience reviewing vertical transactions.  We have a touchstone—namely whether mergers, vertical or otherwise, will result in harm to competition, what GTE Sylvania focused on as interbrand competition, and what Leegin suggested foreclosure could cause.  In conducting that inquiry, we should try to remember what Sherlock Holmes also said: “It is a capital mistake to theorize before one has data.”  Perhaps that is overly strong, and we should say, it would be a mistake to conclude an inquiry based just on theory without a dedicated detective’s desire for detail and data.  A conclusion that economics, evidence and experience suggest, one might say, is Elementary.  Thank you. 

Updated February 10, 2017