In 2012, the New York Stock Exchange changed its policies to prevent brokers from voting uninstructed shares on corporate governance proposals. Although the change was intended to protect investors and improve corporate governance, it has had the opposite effect: a significant number of U.S. public companies are no longer able to amend important parts of their corporate charters, despite the support of their boards of directors and overwhelming majorities of shareholders. Their charters are frozen. My article, Frozen Charters, in the most recent issue of The Yale Journal on Regulation, provides the first empirical and policy analysis of the broker voting change and its significant unintended consequences.

I show that the broker voting change has resulted in the failure of more than fifty charter amendments at U.S. public companies, despite board approval and overwhelming shareholder support, and that hundreds more companies have frozen charters as a result of the change. The rule change has also made it more difficult to amend corporate bylaws and given some insiders a de-facto veto in proxy voting contests. These costs substantially outweigh the negligible benefits of the broker voting change. I compare a number of solutions to address these problems and identify several that would be preferable to the current approach.

The great majority of investors hold shares through brokers or other intermediaries, and many of those investors do not instruct brokers how to vote their shares at annual meetings. Uninstructed broker votes therefore represent a substantial proportion of the outstanding shares of many corporations (an average of 10% in my sample). Since brokers do not have an economic interest in the corporation, uninstructed broker votes could potentially distort voting outcomes. The New York Stock Exchange (NYSE), whose rules effectively govern broker voting, has therefore progressively limited the instances in which brokers may vote without instructions. In 2012, the NYSE promulgated Information Memorandum 12-4 (hereinafter the broker voting change), which restricted brokers from voting uninstructed shares on charter and bylaw amendments.

Although the intention of the broker voting change was to protect investors and enfranchise shareholders, it has done neither. By preventing uninstructed broker voting, the rule change was intended to prevent voting distortions. However, the rule change has instead created a different kind of distortion. Preventing brokers from voting uninstructed shares means that none of those shares will be voted in favor of the proposal, even though some of the shares have beneficial owners that are in favor of the proposal. Where the proposal failed, but would have passed had the uninstructed shares that supported the proposal voted, the outcome is what I term a “distorted fail.”

The main type of distorted fail is frozen charters. A number of corporations with high supermajority requirements for amending parts of their charters are unable to reach these thresholds without uninstructed broker votes. Overall, the broker voting change has significantly increased the proportion of charter amendments that fail despite receiving overwhelming shareholder support, and has increased the likelihood of charter amendments failing in a statistically and economically significant manner. In the three years after the broker voting change took effect, 54 of the 63 companies where charter amendments failed despite receiving overwhelming shareholder support would have had their amendments pass had the broker voting change not been implemented.

These unintended, shareholder-harming consequences should be weighed against the intended consequence of the broker voting change, preventing what I term “distorted passes” from uninstructed broker voting. Uninstructed broker votes will result in a distorted pass when a majority of the shareholders of the corporation prefer that a proposal fail, but uninstructed broker votes in favor of the proposal cause it to pass. However, this almost never happens. Most charter amendments are strongly supported by shareholders, and brokers generally follow management recommendations by supporting these charter amendments. There have been no charter amendments since the broker voting change came into effect that would have had distorted pass outcomes had broker votes been permitted, and only one other management proposal (0.12% of all management proposals during that period).

The foregoing analysis significantly underestimates the true effect of the broker voting change, because only a small number of charter amendments go to a vote each year. I use a novel method to estimate the number of companies that are affected by the consequences of the broker voting change. I estimate that between 13% and 15% of U.S. corporations have been rendered unable to amend part of their charters as a result of the broker voting change. In contrast, I estimate that the broker voting change would prevent distorted pass results at only 0.1% of companies. In short, the broker voting change was implemented to solve a problem that—if it existed at all—had negligible effects.

The implicit goals of the broker voting change were to protect investors, enfranchise shareholders, and improve corporate governance and accountability. However, the substantial negative effects of the broker voting change in creating frozen charters and other distorted fail results significantly outweigh its very limited benefits in preventing distorted pass results. It has clearly failed.

The broker voting change failed because charter amendment votes differ from the other types of votes where unistructed broker voting had previously been disallowed. First, as opposed to director election and executive compensation votes, where shareholders may vote to express their dissatisfaction, shareholders generally support charter amendments, as directors generally do not put forward charter amendments that are unlikely to garner such support. Second, director elections and the passage of executive compensation proposals depends on their support as a proportion of the votes cast at the meeting. However, charter amendments require a proportion of shares outstanding in order to pass. As a result, excluding uninstructed broker votes has a much more significant effect in a charter amendment proposal.

I propose a number of potential solutions, which offer the possibility of reducing the likelihood of distorted fail results, while also minimizing the possibility of distorted pass results. At the very least, and as an interim measure, the broker voting change should be reversed. The NYSE and Securities and Exchange Commission should then work together to develop a policy for the treatment of broker votes on corporate governance proposals that reduces both kinds of potential distortion. The most promising potential solutions appear to be either proportional voting or defining a set of corporate governance matters on which brokers could and could not vote. In this way, the investor protection rationale of broker voting reform could be upheld.

Scott Hirst is a Lecturer on Law at Harvard Law School and an Associate Director of the Harvard Law School Program on Corporate Governance and the Harvard Law School Program on Institutional Investors.