Policy Implications of the Common Ownership Debate

Intervene or Wait for More Research?

In arguing for legal intervention now, Elhauge trenchantly invokes tobacco regulation, noting that tobacco companies staved off life-saving regulation of tobacco products by insisting that scientists supplement the well-understood correlations between tobacco use and lung cancer with scientific “proof” of causal mechanisms. ⁷ Today’s example is resistance to climate regulation by skeptics who claim that causation has not yet been established. As Schmalz points out, science cannot prove causation but can only infer it by ruling out increasingly remote alternative theoretical hypotheses. This process has no logical end, which means an overly rigorous insistence on scientific proof would shut down all forms of regulation.

We have no choice but to evaluate the existing studies and make a judgment. While not everyone has been persuaded by the AST paper, ⁸ there are several facts that are mostly undisputed. First, there is an incentive—a very large one, from the standpoint of theory, given that the law blocks off other forms of collusion—for someone to buy up large stakes in competitors and discourage them from competing with each other. Second, stakes large enough to give influence have already been obtained by major institutional investors and large activist investors as well. Third, there is substantial evidence that common ownership has some effect on the behavior of portfolio firms—with the major discussion about whether this effect is negative or positive, systemic or sector-specific, and large or small. ⁹ Fourth, the spectacular growth of common ownership over the last few decades confronts us with a novel financial structure that concentrates economic power in a handful of firms, creating, at the least, significant potential for great harm. This trend will not end anytime soon. Bebchuk and Hirst predict that the Big Three institutional investors will cast 40% of votes at S&P 500 firms by 2039, up from 25% today. ¹⁰

Finally, the traditional skeptical view of antitrust law associated with the Chicago school, whose echoes can be heard in the argument of the critics, is not looking so good these days. Chicago Schoolists believed that market incentives undermined collusion and monopoly, and antitrust intervention cannot handle the complexity of market behavior outside obvious cases. In taking a hammer to the more restrictive antitrust doctrine that prevailed before the 1970s, they promised that relaxation of antitrust law would not accelerate consolidation. And yet just the opposite happened. ¹¹ All of these developments taken together would normally be considered ample justification for a policy response—at a minimum, investigation, information-gathering, and readiness to intervene, but potentially much more.

The Nature of the Mechanism

The mechanism debate addresses the causal pathway that connects common ownership with higher prices being charged by portfolio firms. The debate initially arose when skeptics asked why they should believe the AST results when it was possible that the correlations between ticket prices and common ownership might reflect an innocent omitted variable. But the debate is important for a second reason. If AST’s empirical results are correct, we need to understand the mechanism, so that we can design a policy response that is suited to the source of the anticompetitive outcomes. Various authors, including AST and Elhauge, ¹² discuss a variety of possible mechanisms. For simplicity, and at the risk of ignoring important complications and variations, I will group them into two categories.

Pecuniary Incentives. Common shareholders can implement compensation packages that discourage managers from competing. In the model of Antón and his coauthors (AEGS), the majority or largest owner offers an incentive contract to each manager of a portfolio firm. ¹³ The model assumes that management seeks blunt incentives because of its own risk aversion and that the compensation package is a compromise between management and the shareholders. As common owners gain a larger share of a firm, the share of the undiversified owners would normally decline. That means the undiversified owners have less influence over management, which can thus implement blunter compensation incentives than the undiversified owners prefer. One can think of this effect as one of crowding-out: as common ownership expands as a result of the demand by individuals for diversification, large undiversified owners like activist hedge funds lose influence. ¹⁴ Note a twist, however. When a common owner increases in size, the anticompetitive effect occurs only if the common owner gains relative to large undiversified shareholders. If large undiversified shareholders also grow, the growth of common ownership may result in unchanged or lower, not higher, markups. The reason is that in AEGS’s model, the performance sensitivity of management compensation depends crucially on the size and influence of large undiversified shareholders relative to other shareholders. While the empirical likelihood of this effect may be minimal under current conditions, it is worth keeping in mind.

Career Incentives and Selection. Common shareholders can also wield their threat to vote in ways adverse to managers’ interests or (in the case of active managers) to sell their stakes. With this threat in place, common shareholders can communicate with managers, secretly or in coded language, directing them to lower output and raise prices. Or common owners can select or influence the selection of managers based on their propensity to compete. They can also influence the selection of directors and use that influence to obtain seats for directors who are loyal to common owners. Here, again, there is a question of how obviously a common shareholder would wield this power. Many commentators doubt that a common shareholder would explicitly instruct managers to collude with the managers of competing firms—a clear antitrust violation. The response is that common shareholders could direct managers to reduce output or raise prices, for which there is some anecdotal evidence. ¹⁵ These instructions do not violate antitrust law in a clear way; a plaintiff would need to prove that the purpose and effect of the instructions were to reduce competition. At the same time, a savvy manager is likely to surmise that a common owner who requests reduction in output may be conveying a similar message to competitors and act accordingly.

These two types of mechanisms can be understood in a crosscutting way in terms of activity or passivity on the part of the common owner. This distinction gets to the heart of the skepticism of the critics who doubt that common owners have the incentive to design an optimal incentive contract that would cause managers to soften competition, especially given the risk that an incentive contract that discouraged competition by blunting incentives would also discourage managers from engaging in effort to cut costs and innovate. ¹⁶ AEGS point out that blunter incentives both reduce incentives to cut cost and to compete, and from the common owner’s standpoint, the gain from less competition may exceed the cost from less effort—especially as the underlying product market becomes more concentrated. Thus, even a relatively passive common owner—one who more or less deferred to management in all respects, normally voting with them when they proposed self-interested blunt incentives in their compensation packages—would produce the common ownership effect of less competition in product markets. But because managers will tend to propose incentives that are too blunt, common owners may, on the margin, intervene with more active measures—like threats to fire managers or attempts to negotiate compensation packages directly. ¹⁷

If either of these mechanisms are in play, what is the right policy response? There are a number of possibilities.

Antitrust Enforcement. Elhauge argues that firms that have obtained large stakes in competitors in a concentrated market violate section 7 of the Clayton Act when their acquisitions have anticompetitive effects. ¹⁸ The argument is simple and intuitive. If AST is correct, then a series of acquisitions by the institutional investors reduced competition and increased prices. Injured private parties—buyers of airline tickets or others—should bring a case, as should the government enforcement agencies.

But there are reasons for thinking that traditional antitrust enforcement will not be adequate to the problem. One is the novelty of the factual setting. Courts have not adjudicated a claim that a firm has caused anticompetitive harm to a market by gradually acquiring a large minority ownership stake in multiple competing firms over decades. And while an anticompetitive purpose is not strictly required, courts may be reluctant to recognize liability on the part of firms that obtained their large stakes by offering index funds, which are widely regarded as socially beneficial. The growth of institutional investors, and their possibly anticompetitive influence on portfolio firms, have the feel of a systemic problem with how markets work rather than the kind of one-off antitrust violation that courts are accustomed to handle.

In 1963, the Supreme Court recognized that courts may block, and break up, mergers that threatened excessive concentration. ¹⁹ The logic straightforwardly applies to common ownership, as Elhauge observes. ²⁰ But there is little enthusiasm for this approach among courts today. This seems to be less about doctrine than the attitudes of judges, who are friendlier to big business than they used to be. While there may be good cases against hedge funds, private equity firms, and other activist financial institutions that have clearly bought up large stakes in competing firms in order to soften competition, creative and ambitious claims against the big institutional investors whose impact on markets through the largely passive accumulation of shares for indexing purpose may be a hard sell.

Finally, there are a range of logistical complexities and problems that may deter antitrust lawsuits. The empirical analysis that litigators would need to undertake in order to prevail is difficult and expensive; there remain a range of methodological disputes that heighten the risk of litigation; judges may regard remedies like divestiture as disruptive; and they also may simply not be able to understand the theory. While, as Elhauge notes, acquisition and even retention of stock may qualify as an anticompetitive act under blackletter antitrust law, ²¹ I suspect that some judges will be bothered by the mechanism issue, just like AST’s critics. They may worry that institutional investors can become liable under the AST theory as a result of actions that they may not even be aware of (say, the sale of stock to the dispersed market by a large undiversified shareholder ²² ). They will not believe that institutional investors order portfolio firms to soften competition unless there is proof of communications; they may not regard voting for a management compensation package that provides management with incentives that are more blunt than socially optimal as sufficiently clear; and they will likely regard the AEGS crowding-out mechanism as a problem for the legislature rather than the courts.

AEGS suggests that the ultimate problem for competition is not so much the rise of common ownership as the decline of the large undiversified investor. In AEGS’s model, competition is retarded even in a Berle-Means word of dispersed shareholders if large blocks are not held by undiversified owners. That seems more like a regulation-of-capital-markets problem than a problem for which traditional antitrust legal principles are suited. And a regulatory approach may seem more appropriate because courts might worry that if they recognize antitrust liability in one case, thereafter potential defendants will not know whether a range of possibly reasonable measures—like mirror-voting, or transferring voting authority from fund families to funds, or limiting their acquisitions to a relatively small percentage of a firm’s stock—will suffice to avoid liability. To be sure, creative and ambitious courts may rule otherwise, but in the meantime, these risks seem likely to deter litigation and certainly large-scale litigation that would provide optimal incentives to minimize the common ownership effect where harmful.

Regulation of Corporate Governance: A paradox lurks in the problem of common ownership. Common owners have a fiduciary duty to exert pressure on managers so that they act in the interest of clients. But to fulfill this duty, common owners need a means of control. And if they have a means of control, they may use it to reduce competition, indeed, in the service of their clients.

We can cut the Gordian knot with a regulatory approach that forbids large common owners to communicate with, or exert control over, managers in any way. The rule would provide that only undiversified owners may communicate with managers. Common owners must remain mum and either not vote or mirror-vote. The effect of this rule would be to create a kind of dual-class ownership structure. Undiversified shareholders would retain voting and communication rights; diversified shareholders would be passive, protected only by their right to dividends and liquidation value.

One virtue of this approach is that it is consistent with one of the original stated justifications of index investing: to minimize expenses. Index funds distinguished themselves from active funds in just this way: because they did not pick stocks, they could refrain from monitoring firms, and of course knowing nothing about the firms, they had little reason to vote. Unfortunately, this justification was in tension with their fiduciary duty to their clients. The reform would require redefinition of the fiduciary duty as requiring the common owner to advance the interests of clients in an administrative capacity only.

A problem with this approach is that it casts its net too large, depriving diversified owners of their control rights even if they are too small to reduce competition. In aggregate, this might leave shareholders at the mercy of managers who do not have to face adverse votes from the bulk of their shareholders. Moreover, in the AEGS model, a common owner may push management of a portfolio firm to compete more rather than less vigorously. If the company has a large undiversified shareholder, AEGS predicts that the common owner will push management to compete less vigorously. But if the company has no large undiversified shareholder and only small shareholders, then the common owner could push management to compete more vigorously. Blocking communication by common owners may on net be unwise.

There are other possible corporate governance reforms, for example, strengthened enforcement of restrictions on interlocking directorates, as it appears that common owners place directors on the boards of competing firms, and may exercise influence through those directors. ²³ However, this reform would leave numerous other pathways of influence.

Regulation of Compensation of Management of Portfolio Firms. One can also imagine a more aggressive approach to corporate governance that regulates management compensation. AEGS finds that common ownership leads to suboptimal (from a social perspective) performance-insensitive management compensation, which in turn results in less competition and higher prices. The natural regulatory response would be to require a higher degree of performance-sensitivity in management contracts. Imagine a law that provides that when common ownership of a firm exceeds a threshold, the compensation plan must be reviewed by the Securities and Exchange Commission. The SEC would have no authority to reject the plan based on the magnitude of compensation (which is the normal focus of complaints about inadequate corporate governance). But it would be permitted to reject the plan if it is insufficiently performance-sensitive.

The standard for judging performance-sensitivity would be based on the compensation practices of firms in other sectors that are more competitive (either because common ownership is less or the product market contains more firms) or firms in the same sector that do not have significant common owners. While government regulation of the terms of employment agreements is generally frowned upon, it may well be justified in this context. It could be modeled on actions against investment advisors for excessive fees under section 36(b) of the Investment Company Act of 1940, which regulates compensation of investment advisors for managing investment funds.

Regulation of Market Structure: A more radical regulatory approach is to limit the size of common shareholders. Imagine that a common shareholder would not be permitted to own more than a small share (say, 1%) of competing firms in a concentrated market; alternatively, a common shareholder could own as large a share of a firm as it wants as long as it owns no shares in its competitors. On this approach, the institutional investor market would presumably segment. There would be behemoths that own, say, 10%–20% of one airline, one oil company, one automobile company, and so on; and there would be minnows that own 1% of all airlines, all oil companies, all automobile companies, and so on. Other investors, some active and others passive, may own more or less large stakes and more or less diversified holdings. Individuals seeking their optimal risk-return trade-off would face a more complex set of options than they do today, but most people would be adequately served by a behemoth (they give up a small amount of diversification in return for simplicity) or a minnow (they obtain maximum diversification but have to wade through multiple options) or a combination (they obtain optimal diversification but must transact with multiple suppliers).

The advantage of this approach is that it is unlikely that diversification exhibits the scale economies that, say, a complex manufacturing operation does. The stock prices of firms within a single sector are highly correlated. The diversification gains from owning all airlines rather than one are small. And because individuals would remain free to buy shares in the minnows, or to buy shares in multiple large funds, those who insisted on maximum diversification would remain free to obtain it at low cost through some trouble.

In an earlier paper, I and my coauthors argued that institutional investors should be given a choice between one option or the other—the minnow approach of complete diversification or the behemoth approach where they sacrifice some diversification but can be as large as they want. ²⁴ (They would also be permitted to be a fully diversified behemoth—with large shares in even competing firms—if they commit not to exert influence by word or vote, as in #3 above.) This regulatory approach would be a lot simpler than incremental remediation through litigation under the antitrust laws. Which is not to say that the regulatory approach would be simple. Regulators would have to define markets and figure out how to address cases where a single firm operates in multiple overlapping markets.

Moreover, market structure regulation is not as simple as converting behemoths into minnows. To take an extreme case, if ninety-nine common owners each owns a 1% stake in every firm in a sector, and (as in AEGS) they do nothing while managers have the remaining stake (or some combination of managers and undiversified funds), the common ownership effect will remain: executive compensation will be insufficiently performance-insensitive, so the firms will be both poorly run and uncompetitive. Market structure regulation must ensure that undiversified shareholders retain a sizable chunk of every firm. The rule allowing common owners to own stakes larger than 1% if they own no more than one firm in a sector would spur some institutional investors to become large owners that are undiversified within a sector, but it might not suffice. ²⁵

If we take the AEGS model as our guide, then policy should be oriented to ensuring that all firms have large undiversified shareholders. It is unclear what the right policy instrument for such a goal would be, but presumably a literal or effective Pigouvian tax on (within-sector) diversification and (in general) smallness could be appropriate. On the other hand, we might worry that extremely large common owners would collude with each other or engage in parallelism, and if this is a case then the expansion of common ownership as such should be the focus of policy.

Stricter Antitrust Enforcement of Portfolio Firms. Theory and evidence say that common ownership’s effect on the competition of underlying portfolio firms is in part a function of the competitiveness of the markets in which those firms operate. ²⁶ If you hold constant the degree of common ownership of the airline industry, and add more airlines, ticket prices should decline. This suggests that an alternative or (more likely) supplemental policy instrument for addressing common ownership is greater antitrust enforcement (or regulation) of the underlying markets. A simple approach would be to lower the HHI thresholds in the Horizontal Merger Guidelines, so that fewer mergers occur and industries remain more competitive. Courts and antitrust agencies could also pay greater attention to efforts by common owners to orchestrate cooperative arrangements among competing portfolio firms. Where common owners call meetings between such firms, offer general directions to them, and so on, ²⁷ the existence of a common owner with large stakes in the competing firms could be a “plus factor” for purposes of section 1 analysis where collusion can be inferred in the absence of a documented agreement.

This approach recognizes that the existing Merger Guidelines disregard a negative externality from mergers—that the incremental concentration of the market caused by a merger will increase the attractiveness of stakes in the firms in the industry for common owners, giving them an incentive to acquire stakes or increase their stakes, and then use their power to further limit competition between the merged firm and its remaining competitors. Or, to put the point differently: we should be willing to give up some scale economies in order to reduce the risk that competition will be excessively reduced in the future through the instrument of common ownership. Merger review currently recognizes that partial acquisitions (i.e., where one firm obtains a minority stake in its competitors) can reduce competition; ²⁸ this logic extends straightforwardly to cases where a third-party investor has partial stakes in the two competing firms.

The problem here is that U.S. product markets are already highly concentrated; it may be too late for an aggressive merger policy to make much of a different, at least in the short to medium term. And if economies of scale in the product market favor large firms, then a more aggressive merger policy may on net cause harm.

The Indirect Costs of Intervention

Several articles written after AST identify a possible benefit of common ownership aside from its facilitation of diversification: common owners may compel portfolio firms to cooperate in socially beneficial ways. We can identify two versions of this argument. First, imagine that Firm A and Firm B are independently developing COVID-19 vaccines. Each firm has proprietary information that it refuses to share with its competitor. A common owner is indifferent as to whether Firm A or B wins the race and so directs the two firms to share or disclose their trade secrets. Here, the common owner enables each firm to capture positive externalities from the behavior of other firms. ²⁹

Second, imagine now that Firm A and Firm B belong to different sectors. Firm A emits greenhouse gases or some other pollutant, and those emissions harm Firm B. A common owner of Firm A and Firm B will direct Firm A to mitigate its greenhouse gas emissions. Of course, the common owner will not direct socially optimal emissions but it will still benefit society. ³⁰ If either of these hypotheses is true, then restrictions on common ownership may be socially harmful.

From both an antitrust and regulatory perspective, however, there is less to these arguments than meet the eye. In antitrust cases, defendants can often argue that a business efficiency justifies behavior that would otherwise seem anticompetitive. Indeed, the research-sharing justification is not much different from the argument that common owners benefit their clients by giving them access to easy diversification. If research efficiencies arise, then in suitable cases, defendants may argue that common ownership (as well as mergers and other forms of cooperation) may be lawful.

The same point can be made about regulation. Simple regulations designed to limit common ownership or control by common owners of portfolio forms may (as so often happens) be required to yield to a more complex approach that creates safe harbors for socially beneficial cooperation across firms. Antitrust law and related regulations that encourage competition are already full of such exceptions.

Finally, regulators should be careful to distinguish diversification that involves common ownership within industries and diversification that involves common ownership (only) across industries. In the second example, a regulation that limits common ownership within sectors will still preserve at least a portion of the socially beneficial behavior: the common owner will restrict greenhouse gases in order to benefit the firms it owns outside the sector of the polluter, while it will have no incentive to reduce competition because it owns only one firm per sector. The first example is more difficult. A regulation that limits the common owner’s ability to own multiple firms in one sector will also prevent it from capturing the externality.

Conclusion

The debate about common ownership echoes the old debate about management incentives in the Berle–Means firm. When shareholders are dispersed, management can choose blunt incentives because no one monitors it. When common shareholders predominate, management can choose blunt incentives because the common owners benefit from them. True, in the AEGS model, common ownership produces a better outcome than dispersed ownership because the common owner cares more about corporate outcomes than dispersed shareholders do and will push for some performance-sensitivity in order to induce managers to cut costs. But the outcome is far from optimal. And it may be worse. With today’s high product market concentration, blunt management incentives imply not just traditional low effort (failure to cut costs and innovate) but also anticompetitive outcomes. The blunt incentives preferred by managers will lead them to compete less as well as innovate less and that matters more when the market is concentrated than when many firms exist.

The biggest common owners also have achieved a level of financial power not seen since the Gilded Age, raising questions about whether they may use that power to achieve political ends or to collude with each other. These risks are independent of the common ownership effect studied by AST and remain empirically unstudied, ³¹ but are just as significant. As the firms continue their inexorable rise in power and influence, we can expect that, like the tech industry before it, their currently benign reputation will give way to public uneasiness and, sooner or later, a government crackdown.