Seven years ago, Einer Elhauge published a call to arms. In a provocative essay in the Harvard Law Review, he urged the antitrust agencies to bring enforcement actions against what he called horizontal shareholding and what we now call common ownership. Common ownership raises antitrust concerns because investors own shares in two or more competitors. While the investors do not control any of the competitors, their joint ownership may be sufficient to cause the competitors to raise prices or otherwise compete less aggressively.
Elhauge’s most powerful argument was that empirical evidence confirmed the hypothesized effect. In two elaborate studies, Jose Azar and co-authors found that increases in common ownership were associated with significantly higher prices in both the airline industry and the banking industry. Given this evidence and Elhauge’s endorsement, other scholars soon wrote supporting articles. Herb Hovenkamp and Fiona Scott Morton, Fiona Scott Morton, Eric Posner and E. Glenn Weyl, were among them.
This initial enthusiasm did not, however, lead to action. There have been no cases and there is no enforcement program. The Department of Justice’s and Federal Trade Commission’s proposed Merger Guidelines do mention common ownership and state that the enforcement agencies have “concerns” with it. But the draft Guidelines do not analyze common ownership in any detail. They do not explain how it might cause anticompetitive effects and what its procompetitive justifications might be. They do not outline any circumstances in which the agencies might challenge common ownership.
This essay suggests that the enforcement agencies ought to take a more muscular approach to common ownership. The Guidelines ought to give it a higher priority and identify the kinds of evidence that might lead to a lawsuit. In what follows, I explain what deflated the initial proposals, argue that those considerations no longer justify the near complete abandonment of interest in common ownership, and outline the evidence that could support a test case.
The Initial Attacks
Two principal arguments derailed the initial proposals. First, critics claimed that the empirical support for the theory was thin and flawed. And initially, the critics were right about the limited support: it consisted of just two studies—the Azar papers mentioned above. Moreover, methodological issues were raised about both studies.
The second attack was perhaps more devastating. Critics argued that no one had explained how common ownership could lead to higher prices or other anticompetitive effects. Of course, a common owner could orchestrate a cartel among the firms whose shares it held. But short of outright collusion—and no one had found evidence of outright collusion—how could this happen? What were the causal mechanisms?
Scott Hemphill and Marcel Kahan analyzed a range of potential mechanisms and concluded that all were either implausible or untested. Lucian Bebchuk, Alma Cohen & Scott Hirst asserted that big index funds charge such low fees that they would not gain any meaningful revenue by pressuring firms to adopt less competitive strategies. Douglas Ginsburg and Keith Klovers stressed that big funds hold shares not only in competitors, but also in vertically related firms, and those vertical investments would undercut their incentive to reduce competition in the relevant market.
These two attacks—on econometrics and governance—sapped the momentum of the initial proposals. Neither the Department of Justice nor the Federal Trade Commission decided to confront common ownership.
The Response
Seven years later, however, the grounds for devoting attention to common ownership are much stronger. There are now fifteen studies that find that higher levels of common ownership are associated with higher prices. Moreover, according to Elhauge, “only two of these empirical studies have been disputed, and the critiques of those two empirical studies have been rebutted at length.” Both the sheer number of studies and their improved methodologies suggest that the evidentiary basis for a challenge to common ownership may now be adequate.
Meanwhile, the corporate governance assault on the theory no longer appears to be so devastating. For one thing, the concern with vertical investments appears to be overstated. It is not clear that funds are as heavily invested in vertically related firms as they are in horizontal competitors. And even if they were, the funds would then be common owners upstream as well as downstream, which would heighten their ability to extract a supracompetitive return from the entire vertical chain.
Second, while index funds do earn small percentage fees, the costs of restricting competition among the firms they hold may be even lower. For example, when funds vote their shares, it costs no more to vote against directors who favor aggressive competition than to vote in their favor.
Third, it now appears that there are a variety of tactics that common owners could plausibly employ to transmit their interest in reduced competition. For instance, many funds regularly communicate with the managements of the firms they hold. They could use those opportunities to press for less discounting, less investment in new capacity, and more emphasis on compensation structures based on industry profits rather than firm-specific profits.
Likewise, funds could withhold their votes when firms or Board candidates propose strategies likely to disrupt the industry consensus. They could vote against hedge funds whose aim is to force the firm to compete more directly against rivals. As Elhauge describes:
[I]n 2015, there was a control contest over management of DuPont, whose main competitor was Monsanto. The fifth largest shareholder of DuPont, the Trian Fund, had no significant shareholdings in Monsanto and launched a control contest designed to replace Dupont’s managers with managers who would behave more competitively against Monsanto. This control contest failed, with the decisive votes to defeat it being cast by the top four shareholders of DuPont (Vanguard, BlackRock, State Street, and Capital Research), who were horizontal shareholders whose financial stake in Monsanto was about twice as high as their financial stake in DuPont.
And active funds (as opposed to index funds) could sell their shares when management embarks on an aggressive campaign to take sales from competitors.
In short, over the last seven years, the case for challenging common ownership has grown. There is much more empirical evidence of adverse effects and significantly greater reason to believe that investment funds can induce corporate officers and directors to curtail their competitive zeal. These developments call for a more active approach to common ownership in the Guidelines.
The Draft Merger Guidelines
The Guidelines ought to analyze common ownership in more depth and explain when it might be challenged. The analysis is straightforward. There is now a substantial literature on the competitive concerns with common ownership, the empirical evidence supporting those concerns, the potential benefits of common ownership, and the hurdles that may prevent common owners from influencing corporate management. The Guidelines can easily describe the major elements of the analysis.
The Guidelines should also identify the kinds of evidence that may warrant a challenge. Three categories seem especially appropriate. First, the relevant market—the market in which the commonly owned firms compete—should be highly concentrated. The agencies can use the same HHI threshold they employ elsewhere in the Guidelines to denote a highly concentrated market (1800).
Second, the level of common ownership should be substantial. The econometric studies measure common ownership by MHHI and the Guidelines should use the level of MHHI that the studies find is likely to be associated with higher prices or other anticompetitive harm.
Third, there should be direct evidence of anticompetitive effects. One could argue, as does Elhauge, that there is no need for direct evidence. The structural evidence (high HHI and substantial MHHI) should be sufficient. After all, fifteen studies have found an association between structural evidence and anticompetitive effects. But the first case challenging common ownership would be a test case and a test case is more likely to succeed—a skeptical court is more likely to accept a novel theory—if there is some direct evidence of harm.
The government could present evidence that in the relevant market higher levels of common ownership are associated with elevated prices, reduced innovation, or other anticompetitive consequences. That evidence could come from an empirical study, a company document, or a journalistic investigation. But given the number of supporting studies that already exist, such direct evidence of impact may not be necessary. What may be essential, in the first case at least, is evidence that a common owner took an affirmative step to dampen competition, such as a direct communication between the owner and an executive, a comment on an earnings call, or a vote against increased competition. The government ought to offer, if possible, a direct connection between common ownership and anticompetitive harm.
If a test case succeeds, the agencies may pursue an enforcement program against common ownership. At that point, the agencies ought to give guidance to investment funds on the relevant markets that are of greatest concern. Eric Posner and co-authors have proposed a method for providing such notice. At this point, however, it is more important that the Guidelines assign a higher priority to common ownership and describe the circumstances that are most likely to result in an action.
John B. Kirkwood is a Professor at Seattle University School of Law and a member of the American Law Institute.