Phantom of the Opera: ETF Shorting and Shareholder Voting

Oğuzhan Karakaş is Senior Lecturer in Finance at the University of Cambridge Judge Business School. This post is based on a recent paper authored by Mr. Karakas; Richard B. Evans, Associate Professor of Business Administration and Donald McLean Wilkinson Research Chair in Business Administration at the University of Virginia Darden School of Business; Rabih Moussawi, Associate Professor of Finance & Real Estate at Villanova School of Business; and Michael Young, Assistant Professor of Finance at the University of Missouri Robert Trulaske, Sr. College of Business.

In our recent paper, Phantom of the Opera: ETF Shorting and Shareholder Voting, we analyze the impact of the short-selling of exchange traded funds (ETFs) on shareholder voting of the underlying securities. We introduce a novel measure of the wedge created between the economic ETF ownership and the voting rights of ETF underlying shares, which we call “phantom shares”. We examine the implications of these phantom shares on the voting process, voting outcomes, voting rights premia, and merger returns. We find that phantom shares, stemming from short-selling of ETF shares (for ETF market making, directional, or hedging purposes), lead to sidelined votes during the proxy voting process. This sidelining appears to be due to the underlying shares backing these ETF short positions going unvoted.

Proxy voting is a fundamental mechanism for shareholder ‘voice’ in corporate governance, including shareholder engagement and activism. With the dramatic surge in passively managed assets across the globe, index funds and ETFs play an increasingly important role in proxy voting. While the debate regarding the efficacy of passive investor voting decisions is in its early stages, our paper addresses a more foundational issue: whether or not the shares of stock underlying the ETFs are voted at all.

To be clear, our evidence does not suggest that ETF sponsors (e.g., BlackRock, State Street, Vanguard) do not vote the underlying shares in their ETF portfolios. Instead, (as a presumably unintended consequence of ETF security design) we show that when ETF shares are sold short, constituent companies’ shares that are not held by the ETF sponsor and would otherwise be voted, are held explicitly or effectively as collateral, and as a result, the holder generally abstains from proxy voting. This abstention effectively decouples the cash flow rights from the voting rights of the corresponding ETF share. We refer to these as “phantom ETF shares”, and the underlying shares held as collateral as “phantom shares”. We demonstrate that these phantom shares are associated with decreased proxy voting and increased broker non-votes, voting premium, and value-reducing acquisitions.

Our findings suggest that phantom ETF shares are costly for investors, since they do not convey the complete voting rights to ETF owners, but are sold at the full price of share, which reflects both cash flow rights and voting rights. Phantom shares also seem to create inefficiencies within the voting process by increasing broker non-votes, and decreasing both the shares voted for and the shares voted against in the shareholder meetings. This becomes particularly important in cases with close votes. Further, we find that phantom shares are positively related to the voting premium, particularly during the meetings with contentious votes.

Our results highlight a significant phenomenon, especially given the recent surge of ETFs, and have important policy implications. Specifically, the existence of phantom shares through the short-selling of ETFs creates inefficiencies regarding the exercise of control rights, and in turn negatively effects corporate governance and the market for corporate control. These issues become more crucial during times when the markets are bearish and/or when the votes are critical and very valuable. We believe our findings are particularly important when considered against the simple alternative of investing in index mutual funds, which do not appear to suffer from a similar lack of voting rights.

The US proxy voting system needs to be updated, as various inefficiencies and inaccuracies have been discussed in the corporate law literature. Indeed, the accuracy and transparency of the US proxy voting process has increasingly been under the spotlight of the U.S. Securities and Exchange Commission (SEC). Following up on the SEC’s July 2010 “Concept Release on the U.S. Proxy System” and the November 2018 “Roundtable on the Proxy Process”, which provide the blueprint for the proxy system in the US and discuss “proxy plumbing” problems such as over- and under-voting, the Investor Advisory Committee (IAC) of SEC has recently called for a deeper investigation of the impact of securities lending on voting rights. We believe our paper makes a timely contribution to this inquiry.

Similar to this need for an overhaul of the proxy voting system, there is also a need for a robust regulatory framework for ETFs. We believe our findings may inform these debates and the construction of the related systems/frameworks, given the complexity of the ETF market-making process and potential associated unintended consequences of ETF shorting and collateral management. Existing financial systems/regulations tend to focus more on the cash flow rights and to ignore the other contractual rights of the shareholders, such as right to vote and right to sue. A distributed blockchain ledger may help tracing the shares accurately and efficiently. However, it is not very clear whether this itself would streamline voting, as it may bring other risks such as risks from increased transparency and/or resolve the problem we identify. Further disclosures and research are needed in this area.

We also believe the phantom shares measure we have introduced here may prove helpful to academic researchers as a plausibly exogenous instrument for the change in ownership/control. We have exploited this aspect of phantom shares in our analysis of the acquirer returns but there are numerous other potential applications. A similar approach utilizing phantom shares could be adopted in addressing important and interesting issues in corporate finance/governance.

The full paper is available for download here.

Both comments and trackbacks are currently closed.