How Horizontal Shareholding Harms Our Economy—and Why Antitrust Law Can Fix It

Einer Elhauge is the Petrie Professor of Law at Harvard Law School. This post is based on Professor Elhauge’s recent paper.

Related research from the Program on Corporate Governance includes Horizontal Shareholding (discussed on the Forum here) and New Evidence, Proofs, and Legal Theories on Horizontal Shareholding (discussed on the Forum here), both by Einer Elhauge; The Agency Problems of Institutional Investors by Lucian Bebchuk, Alma Cohen, and Scott Hirst (discussed on the Forum here); and Index Funds and the Future of Corporate Governance: Theory, Evidence, and Policy by Lucian Bebchuk and Scott Hirst (discussed on the forum here).

In my initial Harvard Law Review article on horizontal shareholding, I showed that economic theory and two empirical studies of airline and banking markets indicated that high levels of horizontal shareholding in concentrated product markets can have anticompetitive effects. I argued that those anticompetitive effects could help explain longstanding economics puzzles, including executive compensation methods that inefficiently reward executives for industry performance, the sharp rise in the gap between corporate profits and investment, and the growing increase in economic inequality.

My claims have all been hotly contested. However, as I show in a new paper, new proofs and empirical evidence strongly confirm my economic claims. One new economic proof establishes that, if corporate managers maximize either their expected vote share or re-election odds, they will maximize a weighted average of their shareholders’ profits from all their stockholdings and thus will lessen competition the more that those shareholdings are horizontal. Another new economic proof shows that with horizontal shareholding, corporations maximize their shareholders’ interests by making executive compensation less sensitive to their own firm’s performance because that reduces competition between firms in a way that increases shareholder profits. Neither new proof requires any communication or coordination between different shareholders, between different managers, or between shareholders and managers.

These new economic proofs have been confirmed by two new cross-industry empirical studies and three new market-level studies. One cross-industry study shows that increased horizontal shareholding does make executive compensation less sensitive to their own firm’s performance, just as the economic proof predicts. The other new cross-industry study shows not only that the recent historically large gap between corporate investment and profits is mainly driven by horizontal shareholding levels in concentrated markets, but also that within any industry, the investment-profit gap is mainly driven by those firms with high horizontal shareholding levels. The three new market-level studies find that horizontal shareholding increases seed prices and both reduces and delays competitive entry into pharmaceutical markets.

I further provide new analysis rebutting various critiques of the earlier studies of airline and banking markets. While a few of these critiques are valid, addressing those valid critiques actually increases the estimated price effects. The other critiques are all mistaken. For example, some rest on endogeneity claims that are flatly contradicted by the evidence. Another critique uses purported proxies for horizontal shareholding that are actually negatively correlated with horizontal shareholding and uses market models that wrongly assume longer airline routes have lower costs. Other critiques erroneously measure horizontal shareholdings without aggregating the shares held by the same fund families, ignore actual market shares, exclude the transactions most likely to have price effects, and wrongly set many horizontal shareholding rights to zero.

Nor are the findings of anticompetitive effects undercut by a recent cross-industry study that purports to show that horizontal shareholding has no robust effect on profits or investments.  This study actually finds that large increases in ΔMHHI do increase profits. It finds no statistically significant effect from smaller increases in ΔMHHI, but that is not surprising given that even for horizontal mergers, it takes a ΔHHI of at least 200 to make anticompetitive effects likely. Further, because virtually all of the many variables used in this study depend on industry definitions that do not accurately reflect antitrust markets, all if its regressions suffer from attenuation bias that leads it to underestimate effects. All its regressions also either fail to correct and aggregate the data on horizontal shareholding levels or use control variables that create problems of multicollinearity and reverse causality.

In short, contrary to the claims of some, we do not have the sort of empirical uncertainty that justifies further delaying any enforcement actions against horizontal shareholding. Further, although some argue that the causal mechanisms or horizontal shareholder incentives to create anticompetitive effects are unproven or implausible, I debunk such claims in another recent paper, The Causal Mechanisms of Horizontal Shareholding. Moreover, my proposal is simply that antitrust agencies investigate concentrated markets with high horizontal shareholding to ascertain whether anticompetitive effects exist in those markets, so any empirical uncertainty would be resolved in enforcement actions about specific markets.

My new paper also shows that antitrust law already provides a legal remedy. Whenever horizontal shareholding has anticompetitive effects, it violates Clayton Act §7’s ban on any stock acquisitions that have anticompetitive effects. As I show, this interpretation is dictated by the legislative text, structure, and history and this legal remedy raises no insuperable administrability problems. Any horizontal shareholding that has anticompetitive effects also violates Sherman Act §1. Indeed, the very name of the legal field—antitrust law—comes from the fact that the Sherman Act aimed to prohibit certain pre-1890 trusts that were themselves horizontal shareholders in competing firms. It has thus always been the case that horizontal shareholding by a common shareholder is an agreement or combination covered by Sherman Act §1.

I further show that EU competition law can also tackle horizontal shareholding. Although EU merger control law is narrower than Clayton Act §7, EU law’s prohibition of anticompetitive agreements and concerted practices under EU Treaty Article 101 is at least as broad as Sherman Act §1’s prohibition of anticompetitive agreements, and is thus broad enough to condemn anticompetitive horizontal shareholding. Even broader is EU law on collective dominance and excessive pricing under EU Treaty Article 102, which provides a straightforward solution to the problem of horizontal shareholding.

Finally, I show that even if courts or agencies misinterpret competition law not to apply to horizontal shareholding directly, such horizontal shareholding still alters traditional merger analysis. After all, such traditional analysis requires assessing whether mergers and cross-shareholdings have likely anticompetitive effects, and the likelihood of such effects is increased by horizontal shareholding in concentrated markets. Indeed, the less that our antitrust regimes do to directly tackle horizontal shareholding, the lower the concentration levels they can tolerate when doing traditional merger analysis. Horizontal shareholding can also mean that a merger that would otherwise be deemed non-horizontal (because the merging firms compete in different markets) should instead be deemed horizontal if the merger increases shareholder overlap between the merged firm and its competitors. Given these implications, rising levels of horizontal shareholding, especially if we continue to do nothing to directly tackle them, provide strong support for current antitrust movements that decry our increasing levels of national industrial concentration.

The complete paper is available for download here.

Both comments and trackbacks are currently closed.