Misguided fears of common ownership

2 June 2019

A new analysis suggests that fears about the competition effects of common ownership are overblown.

So-called common ownership – where a fund manager owns shares in a number of competitor firms within the same industry – has grown dramatically over the last three decades. Whereas in 1990 only 17% of S&P500 firms had a blockholder that also owned a block in a competitor firm, this had increased to 82% by the end of 2015.

A competition problem?

The governance community has been set alight by a paper by Jose Azar, Martin Schmalz, and Isabel Tecu that claims to prove that increases in common ownership in the US airline industry caused increases in ticket prices on key routes. The core idea is that shareholders holding stocks across an industry group create anti-competitive incentives that cause industry players to act in a way (particularly through pricing) that increases the share of the economic pie taken by investors at the expense of consumers. If true, this would have significant anti-trust implications. The accusation has excited the imagination of policy makers around the world and has led to some fairly drastic policy proposals including: limiting fund managers to holding at most 1% of any company; limiting the number of competitors that a fund manager can hold shares in within a given industry; disenfranchising index funds; and including common ownership considerations within the remit of competition authorities.

A major cause of the growth in common ownership has been the growth in index investing, in particular the ‘Big Three’ of BlackRock, State Street, and Vanguard, who own stakes of 5% or more in the vast majority of S&P500 firms. The common ownership debate has therefore been a useful vehicle for beating up on index funds, for those who consider the growth of such funds at the expense of their active counterparts to be an existential threat to shareholder capitalism. 

Questionable causation

But a recent working paper by Katharina Lewellen and Michelle Lowry, which Lewellen presented the 3rd BI Corporate Governance Conference in Oslo in May 2019, casts doubt on these claims. With policymakers reaching for their drafting pens, this analysis is important. 

Studies of causality in finance are fraught with technical challenges. Unlike in medical research, it is not possible to create a randomised control trial. Instead, finance experts look for an event that can be considered an exogenous shock that creates ‘treated’ and ‘untreated’ firms that can be compared. The event used by Azar et al was the acquisition of BGI by BlackRock in 2009. As the holdings of the two managers was combined, this gave rise to significant increases in cross ownership in some firms but not in others. By comparing behaviour in firms where cross ownership increased against those where it did not, the authors seek to identify a causal impact of increases in common ownership. Another trick has been to look at companies entering the S&P500 or reconstitution of the Russell indices. As companies come into an index, this results in purchases by index funds, which again can increase common ownership. 

Overall, 14 papers have looked at the common ownership issue in various ways, and 11 claim to find a competition effect from common ownership. Lewellen and Lowry have reviewed these, paying particular attention to the event that is used to identify causality. Given the profile the issue is gaining amongst policy makers, this is an important paper. 

Use of index construction to identify effects is shown to be problematic. As well as leading to increases in cross ownership, entering the index also leads to increases in institutional ownership overall, particularly by index funds, and a reduction in large blockholders, who sell out to the institutions. Untangling all these effects is difficult and likely to render results unreliable. The authors come down firmly in favour of using mergers between financial institutions as the best source of an exogenous shock to levels of common ownership. 

The paper undertakes a detailed analysis of relevant institutional mergers over a c. 20 year period, expanding beyond the BlackRock-BGI merger used by Azar et al. They find that:

• Positive results relating to common ownership are disproportionately driven by the BlackRock-BGI merger. Although this is just one merger, it’s scale makes it responsible for the vast majority of observations of where common ownership in firms increases.

• When the BlackRock-BGI is removed and the analysis restricted to other mergers in the 20 year period, the relationship between common ownership and competition is no longer observed.

• The BlackRock-BGI merger took place in the midst of the financial crisis, which caused differential responses across firms and industries, which the authors find is quite likely to explain the observed changes in competition, rather than common ownership.

Katharina Lewellen was keen to point out in her presentation that lack of evidence does not mean there isn’t an effect. One interesting line of inquiry would be to compare industries that are mainly domestic and where most of the market is made up from listed firms with those industries were there are strong international or private competitors. Any effects from common ownership should be larger in industries with the former characteristics. Where there are strong overseas or private competitors who do not have common ownership with the domestic listed players, any anti-competitive effects should be much harder to sustain, and so we should observe a difference between these types of industry if the common ownership thesis is valid.  

So we should retain an open mind. But the work of Lewellen and Lowry has cast significant doubt on the significance of the findings to date. It must be hoped that this stays the hand of regulators until we’re really clear what problem we’re trying to solve, if there’s a problem at all. 

Indirect effects?

Anyone who has had a stewardship interaction with a major fund manager would raise an eyebrow at the idea of index funds acting as puppeteers, organising industry collusion. Indeed a second recent paper that plays into this debate, by Lucian Bebchuk and Scott Hirst, makes claims (which would certainly be disputed by the funds themselves) that far from being over-involved, index funds are largely absent in providing oversight of companies. They too conclude that the common ownership concern is a ‘red herring’. 

But could there be unintended indirect events of extensive common ownership, especially by index funds? Index funds have been very effective at using their voting muscle to impose market-wide standards of governance in areas such as board composition, executive pay, shareholder rights (including pushing back against dual class shares, staggered boards and the like). But at what point does a market wide standard become an impediment to competition?

One example from the UK is the area of executive pay. Reinforced by proxy agencies, and spurred on by public opinion, the major funds are increasingly boxing companies in on pay. Companies are at risk of significant votes against if they adopt a non-standard pay design or if they increase pay, even for a new recruit. This has in effect taken pay off the table as a mechanism by which firms can compete – instead pay is determined by an increasingly standardised rule-book. If extended to other areas of corporate governance it is possible to see how animal spirits might be subdued and firms lulled into being safe and cosy rather than aggressively competitive. 

Time to take a breath

Some of these indirect effects seem more likely channels for influencing firm behaviour. We should not assume that there will be no effects from the revolution in ownership patterns that we are seeing in capital markets. But overall we need to be very careful before pursuing some of the draconian interventions proposed. Far stronger evidence would be needed. It’s time to take a breath and look at the facts before imposing solutions that could be worse than the problem. Lewellen and Lowry have made an important contribution to inform the debate.

---