Surprise Parties: A huge data set gives Ryan Bubb and Emiliano Catan LLM ’10 new insight into how mutual funds vote their shares

In 2009, Netflix concluded a three-year open competition by awarding $1 million to a team of researchers that had created the best algorithm to predict user ratings for movies. A decade later, when Robert B. McKay Professor of Law Ryan Bubb and Professor Emiliano Catan LLM ’10 set out to achieve a clearer understanding of institutional investors’ voting behavior and corporate governance preferences, they decided to take a page from Netflix’s playbook.

Ryan Bubb
Ryan Bubb

The training data set that Netflix provided to participants in its competition contained more than 100 million ratings. Bubb and Catan, who used mutual fund voting information from the years 2010 through 2015, found themselves grappling with a data set comprising more than 28 million votes, so they employed similar machine learning techniques to crunch an otherwise unwieldy amount of data. They published their findings in “The Party Structure of Mutual Funds” in the June 2022 Review of Financial Studies.

Emiliano Catan
Emiliano Catan LLM ’10

Bubb and Catan approached the task without a specific hypothesis, letting the data establish the narrative. Using dimension reduction techniques and cluster analysis, the two identified three main types of mutual funds in terms of corporate governance philosophies expressed through voting. A portrait emerged of a trio of mutual fund “parties,” which Bubb and Catan call the Traditional Governance Party, the Shareholder Reform Party, and the Shareholder Protest Party. Like members of a political party, members of mutual fund parties tend to vote as a bloc in shareholder elections.

The Traditional Governance Party votes with company management at higher rates than the other two parties, particularly when an issue is traditionally understood as being within the jurisdiction of a board rather than of shareholders. On the other hand, this party is likeliest to oppose management on proposals involving fundamental shareholder rights (typically, initiatives to alter basic corporate governance rules), often coming down on the side of shareholders. The party also frequently disagrees with management in proxy contests.

In contrast, the Shareholder Reform Party votes against management and requests specific corporate governance reforms at much higher rates than its two counterparts. The party supports shareholder intervention in matters traditionally considered within the board’s lane, including corporate social responsibility (CSR) and executive compensation.

In some proposal categories, the Shareholder Protest Party opposes management at a rate between the anti-management voting rates of other parties. But with uncontested director election proposals and non-binding say-on-pay proposals (allowing shareholders to vote on executive compensation), which together constitute more than three-fourths of the sampled votes, the party defies management considerably more often than the other two parties do—essentially engaging in “protest” votes that are largely symbolic.

Bubb and Catan suggest that their novel concept of a three-party mutual fund voting structure can shed new light on corporate governance issues. For instance, recent debate on corporate governance has focused on the shift from active to passive fund management among institutional investors and whether that trend is leading to decreased monitoring of corporate management.

The three-party lens, Bubb and Catan assert, can help in making more accurate predictions. The Traditional Governance Party includes the “Big Three” passively managed funds—BlackRock, Vanguard, and State Street—and is the largest party in terms of total assets managed. As Bubb and Catan establish, these passive managers oppose management in votes concerning fundamental shareholder rights even as they defer to management on other matters, making them far more than rubber stamps.

Another common concern involves the growing influence of proxy advisors on corporate governance. But Bubb and Catan argue that the growth of the Big Three passively managed funds is likely to weaken the influence of proxy advisors, whose recommendations those funds tend not to follow. They also find that the recommendations of the two largest proxy advisor firms, ISS and Glass Lewis, correspond to the distinct voting behaviors of the Shareholder Reform Party and the Shareholder Protest Party, respectively—even though ISS and Glass Lewis’s proxy voting guidelines articulate similar corporate governance philosophies. The co-authors find that proxy advisors’ recommendations help define differences in voting behavior among the three parties—while at the same time the advisors’ influence on votes overall appears to be weaker than generally believed.

Bubb explains that their findings create a more complex picture than the more black-and-white takes he and Catan have seen in other analyses of institutional investor voting. “We’re both economists,” says Bubb. “We are reductive, frequently…. And that reductive spirit in fact can be a great strength of economic methodology.

“On the other hand,” Bubb adds, “there are times when the reductive approach elides important distinctions.” Catan calls legal scholars “more attuned to some of the more fine-grained dimensions over which votes might differ,” and says that Bubb and he “were able to bring those nuances to bear in analyzing the data.”

The US Securities and Exchange Commission (SEC) has taken note of Bubb and Catan’s work. In December 2018, Pierrepont Family Professor of Law Robert Jackson Jr., then an SEC commissioner, delivered remarks at a public Federal Trade Commission hearing at NYU Law in which he cited Bubb and Catan’s working paper.

“Large institutions, and especially index investors, wield unprecedented power in corporate elections—power that comes from the savings of millions of American families,” said Jackson. “What can the SEC do to make sure that ordinary investors understand how institutions are wielding their influence? We should start by shining light on how institutions vote Americans’ money.”

Jackson went on to describe the newly identified “party structure” of mutual fund voting as a potential means of elucidating voting for retail investors. “Whether the Bubb-Catan approach, or a competing method, is the best way to show how institutional investors vote is beyond the scope of my testimony, [but] we at the SEC should be taking steps to make sure that American investors understand how their retirement savings are being used in corporate elections,” Jackson said.

The following July, the paper was cited multiple times in the SEC’s rulemaking petition for real-time disclosure of proxy votes. In November 2022, the commission adopted rules to make investment funds’ proxy voting data more user-friendly. Those actions should make it easier to realize Bubb and Catan’s expressed desire at the end of their article: “We hope the introduction of our measures of mutual fund corporate governance preferences to the literature will enable other researchers to test quantitatively a range of theories and hypotheses about corporate governance.”

Posted January 25, 2023