Archives For

[TOTM: The following is part of a symposium by TOTM guests and authors on the 2020 Vertical Merger Guidelines. The entire series of posts is available here.]

This post is authored by Timothy J. Brennan (Professor, Public Policy and Economics, University of Maryland; former Chief Economist, FCC; former economist, DOJ Antitrust Division).

The DOJ Antitrust Division and the FTC have issued draft proposed Vertical Merger Guidelines (VMGs). These have been long-desired in some quarters, and long-dreaded in others. The controversy remains, although those who long desired them may dread what they got, and vice versa. 

I’ve accepted Geoff’s invitation to offer some short thoughts on this blessedly short draft. I’ll focus on what one might call “pure” vertical mergers, as opposed to those based on concerns regarding potential entry or expansion by a vertically integrated firm into its upstream or downstream market. After looking at what the draft classifies as unilateral effects, I’ll then turn briefly to coordinated effects, burden of proof, and the whether we want to go there at all

A notable omission

It’s useful to think of pure vertical mergers mergers in a reverse direction; that is, not the harms they would create, but the benefits of blocking the merger even if it were harmful. The usual argument against prosecuting such mergers is that one would be left with separate firms charging above-competitive prices at one end or the other — and if they do it at both ends, creating double marginalization. 

In that regard, the draft has a notable omission: a skeptical eye toward mergers when a firm with market power is unable to charge a consequently high price. The usual justification is price regulation of a monopoly. Concern about the harm of vertical integration as a means to evade such regulation was the basis for the divestiture by AT&T of its then-regulated, then-monopoly local telephone companies, the mirror image of blocking a vertical merger. This concern also formed the basis of Federal Energy Regulatory Commission orders (here and here) separating control of regulated electricity transmission lines from ownership of unregulated generation.

More controversially, one could apply this in contexts where pricing is limited by something other than explicit regulation. TOTM has recently featured a debate between Geoff Manne and Dirk Auer on one side and Mark Lemley, Doug Melamed, and Steve Salop regarding the FTC’s case against Qualcomm’s licensing practices, where the issue (I think) is about whether FRAND licensing requirements limit Qualcomm’s ability to extract the full rental value of its patents. Perhaps even more out there, I’ve suggested that transaction costs keeping Google from charging directly for search could justify concern with discrimination in favor of its offerings in other markets, without having to reject a consumer welfare standard and where “free” search could justify rather than impede an antitrust case. Whether non-regulatory restrictions on price should justify vertical concerns is, to say the least, complicated. You’ll see that word often, and I’ll come back to it at the end

Raising rivals’ costs and foreclosure: Why would the merger matter?

The seeming truism of raising rivals’ costs is that to raise someone’s costs, you have to raise the price of an input they use or a complement they need. If a merger (or vertical restraint) extends control over an input or complement, the competitive risk arises because that complement market is being monopolized. Accordingly, enforcers need to ensure they are looking at competition and ease of entry in the market that’s being monopolized, not the market in which the bad guy who benefits might operate. For example, in the case involving Intel’s loyalty rebates to computer makers that allegedly excluded AMD’s processing chips, the relevant market was not chips but computers, and it is necessary to explain why AMD couldn’t get new computer makers to use its chips or vertically integrate itself into computers. 

With a “pure” vertical merger, however, the raising rivals’ cost or foreclosure story depends not on the creation or expansion of market power over an input or complement. It must depend on how vertical integration leads to the exercise of power in that market that was not being exercised before. That requires that vertical integration changes the strategic interactions determining price and quality. 

If this is the story, however, the agencies should make clear that they bear an obligation to explain how vertical integration matters, rather than assume a competitive outcome otherwise. The recent (failed) Justice Department challenge to AT&T’s vertical acquisition of Time Warner programming bears this out. DOJ’s core claim was that post-merger, Time Warner program services (HBO, CNN) would strike tougher deals with other video distributors, leading to higher licensing fees, because they would know that a failed negotiation would increase subscribership to AT&T. However, once a program service licenses to any video distributor, it realizes that if other video distributors don’t take their service, it will increase subscribers and profits from its initial deal. If that change in bargaining position benefits whoever gets the first deal, distributors presumably can cut a deal to be the first, without vertical merger being necessary. 

Might vertical merger still matter? Perhaps, but plaintiffs should have to explain what transaction costs prevent such a nominally anticompetitive deal. Critics of indifference to vertical mergers, and the draft VMGs, point out that double marginalization might be eliminated without vertical integration, through use of contracts that include fixed fees and sales at marginal cost. A similar skepticism regarding the need for vertical integration to achieve anticompetitive ends is similarly appropriate. Establishing that need may be, well, complicated

Coordination and symmetry

I have less to say about coordinated effects. Knowing when firms decide to switch from competing to cooperating is probably a matter of psychology and sociology as much as economics. However, economics does suggest that collusion is more likely the more symmetric are the positions of the potential colluders, to help find a mutually agreeable tactic and reduce the need for side payments and the like. To the extent symmetry is relevant, a pure vertical merger will facilitate collusion only if other market participants were similarly integrated. And absent market-wide collusion to integrate, this integration of the other participants could be reasonably interpreted as evidence of efficiencies of vertical integration. This would make a coordinated effects vertical merger case, well, complicated.

The burden of proof

As part of a widely voiced dissatisfaction with the last few decades of antitrust enforcement, some commentators have suggested that vertical mergers deserve no more of a presumption of being good than horizontal mergers. I would argue that the strength of evidence necessary to block a vertical merger should remain greater than that for horizontal mergers. This is not (just) because mergers between complements are more likely to lead prices to fall — eliminating double marginalization being an example — while horizontal mergers are likely to bring about upward pricing pressure. It is that the behavior that warrants concern with horizontal mergers, joint price setting, is illegal. Thus, the horizontal merger might be the only way to enable that conduct. 

Contracting alternatives to vertical merger abound, as two-part pricing to eliminate double marginalization and sequential per-subscriber pricing in video markets illustrate. Absent regulation, the transaction costs necessary to create and exercise market power but for vertical integration will be lower than those necessary to create and exercise that power but for horizontal merger. This difference in transaction costs means that horizontal merger is more likely to be necessary for harm than vertical merger. Thus, proving that the latter are harmful should be, well, more complicated.

So, are pure vertical mergers worth the trouble to prosecute?

Vertical mergers might be problematic, but cases are necessarily going to be complicated. This raises the question of whether the benefits of prosecuting such mergers are worth the costs, at least outside the regulated industry context. These costs are not just the cost of litigation but the costs to companies of having to hire antitrust counsel and consultants to try to determine whether their merger is likely to be challenged, how much negotiation with the agencies will be entailed, and what the likelihood of loss in court will be. Some aspects of antitrust enforcement, primarily determining who competes with whom and how much, are inevitably complicated. But it is not clear that we are better off expending the resources to see whether something is bad, rather than accepting the cost of error from adopting imperfect rules — even rules that imply strict enforcement. Pure vertical merger may be an example of something that we might just want to leave be.

brennanTim Brennan is a professor of public policy and economics at UMBC and a senior fellow with Resources for the Future (RFF).

When I first started working in antitrust at the Justice Department over thirty years ago—there’s a hard reality to accept—the Antitrust Division was then embroiled in an effort to reform the regulation of oil pipelines. The argument on this now obscure issue was that effective prevention of the exercise of market power by natural monopoly pipelines required both clear pricing standards and effective separation of control of the pipeline capacity from the shippers who often owned the pipeline as well. Among other things, this led the Division to take an active role in the proceeding to set the rates to send oil through the Trans-Alaska Pipeline.

This largely forgotten issue had a much more consequential successor—the AT&T divestiture in 1984, settling an antitrust case filed a decade before. That case took the hard line that to prevent anticompetitive abuses, regulated monopolies and competitive services that rely upon them should rest in completely separate companies. Although the 1996 Telecommunications Act superseded that settlement, the principle of separating control of regulated assets from firms that compete in services that use those assets survives. The prominent example is electricity. Many state regulators ordered local distribution utilities to divest power plants, and federal regulators require that power transmission be supplied by regional organizations independent of the control of the competing generators that use them.

When Assistant Attorney General William Baxter announced the AT&T divestiture in 1982, he also announced that he was dropping, with prejudice, the even older antitrust case against IBM. The distinction was that AT&T operated in both regulated and unregulated sectors, while IBM did not. AT&T had an incentive to evade regulatory price constraints by creating an artificial competitive advantage by impeding its competitive market rivals’ access to its local service monopolies (“discrimination”). Moreover, depending on the form that regulation might take, the ability to shift costs from unregulated to regulated sectors may allow cost shifting enable a regulated firm may to make predatory threats credible (“cross-subsidization”). In the old AT&T case, this was called “pricing without regard to cost”.

Mr. Baxter’s distinction rested on the crucial point that regulation is a complement to antitrust, not a substitute. Absent regulation, refusals to deal are not presumptively harmful, and may be beneficial, while predatory threats are notoriously difficult to validate. This is why Mr. Baxter, no great hero to antitrust activists, famously pledged to litigate U.S. v. AT&T “to the eyeballs.” Continue Reading…

brennanTim Brennan is a professor of public policy and economics at UMBC and a senior fellow with Resources for the Future (RFF).

As evidenced by this on-line symposium, the handling of cases under the rubrics “monopolization,” “single firm conduct”, or “abuse of dominance” continues to be debated by the competition policy community. This debate, as evidenced by the Antitrust Division’s Sept. 2008 single firm conduct report and the FTC responses, is not restricted within the U.S. The European Union has published “Guidance Papers” on standards for exclusionary conduct under Article 82, and the Canadian Competition Bureau recently issued draft guidelines for prosecuting conduct under the abuse of dominance provisions of Sec. 79 of its Competition Act.

Almost any significant antitrust case will engender controversy over the facts, e.g., damages resulting from cartel conduct or market definition for mergers. The controversy over single firm conduct runs deeper. Much of this contention arises because the direct focus of the conduct, harm to rivals, is also the byproduct of vigorous competition. Despite everyone having learned to utter the mantra “protect competition, not competitors,” we find a line drawn between two sides. To caricature the bifurcation only slightly, one side, which I’ll call the skeptics, would set the burden of proof very high, with harms to competitors presumptively competitive. The other, here called the activists finds enforcement lax, is more willing to protect competition by protecting competitors.

I propose to resolve the controversy by positing that both sides are right-but within separate categories of monopolization cases. Continue Reading…