The past several months have seen a resurgence in the years-long debate over proxy mechanics. One focus of the debate has been issuers’ concerns about “empty voting” and “over-voting,” which relate to the registration of votes that should not be counted. On the other side of the coin, shareholders have varied concerns about their ability to confirm that their votes have been accurately reflected in the tabulation results.
In response to questions we have received on this subject, we have prepared the below answers in order to better explain the factual background and to provide some insight into these issues. This is the third in a series of advisories focused on issuer-shareholder communications and the related proxy voting system. Prior advisories have focused on the SEC’s NOBO-OBO rules (click here to view), and the SEC’s rules governing the content of shareholder communications (click here to view). In this Q&A we focus on these issues related to proxy voting processing.
What are over-voting and empty voting?
Both “over-voting” and “empty voting” refer to types of errors that can occur in the way that proxy votes are counted. Over-voting refers to a situation where a bank or broker communicates to the vote tabulator more votes than its clients are technically entitled to register. (The tabulator is hired by the issuer to make a final tally of the votes that are submitted.) Empty voting refers to a situation where a shareholder has voting rights in the shares to be voted, but lacks full economic interest in those shares.
Over-voting differs from empty voting in one important respect. While the former is a processing issue that is always—or virtually always—corrected before a final outcome, when the latter occurs, it could undermine the legitimacy of a final outcome.
How does over-voting happen?
Over-voting can occur when a broker’s client believes that he or she has more votes than the client does in actuality. This can happen because—under some circumstances—a client may own more shares than he or she is entitled to vote. The reasons are technical, but the phenomenon most typically occurs when the client has shares held in a “margin account” (a type of account that is created when the client has borrowed money from the broker such that the latter retains a collateral interest in the account). In these situations, the broker is permitted to “lend out” the shares (including their voting rights) to other brokers, even while the shares are still reflected on its client’s account statement. This practice whereby the broker lends out the shares is called “re-hypothecating” the shares.
Thus, for example, assume that Mr. Smith purchases 100 shares of ABC Company on margin through his broker. His broker lends to Mr. Smith some or all of the money that Mr. Smith uses for the purchase, and the client’s new shares are placed in a margin account. The broker subsequently “lends out” 50 of Mr. Smith’s shares. During the pendency of the broker’s “loan,” Mr. Smith is entitled to only 50 votes. Typically, the broker will have loaned the shares to someone who wishes to sell the shares as part of a short sale transaction, so the other 50 votes would belong to the third party who purchased the other 50 shares, who in turn has also now become the record owner of the shares.
Prior to the company’s annual shareholders meeting, Mr. Smith receives a voter instruction form (VIF) reflecting that he has 100 votes, and Mr. Smith casts 100 votes. If Mr. Smith’s broker reports all 100 votes to the tabulator, the extra 50 votes may result in the broker reporting more votes in aggregate than it holds itself or on behalf of its clients. Note, however, that if some other client of the broker declined to vote 50 of his or her shares, Mr. Smith’s 100-share vote may not in fact result in over-voting.
Does over-voting sometimes skew elections in the sense that it results in the registration of too many votes?
No. If a broker reports too many votes in aggregate, the tabulator will notify the broker of the discrepancy. The broker then rectifies the problem, and resubmits its voting report. How does the tabulator know that the broker has reported too many votes? All transfers are netted at the level of the depositories, such as DTCC, which notifies the tabulator of the number of shares a particular broker actually holds.
Automated procedures have been in place for years to eliminate virtually all instances of over-reporting so that does not carry though to a final vote tally. In this respect, at least, over-voting is arguably no longer a problem.
Why can’t the broker avoid reporting too many votes in the first place?
Brokers can avoid over-voting, and today they generally do avoid it. There are two approaches: “post-reconciliation” and “pre-reconciliation.”
A broker following a post-reconciliation model allows its clients to vote all the shares that they hold in their accounts, including any shares that may have been re-hypothecated. If the broker subsequently determines that the process will result in more aggregate votes than it is entitled to register, it will reduce votes in some order of priority, generally starting with re-hypothecated shares in margin accounts and its own proprietary shares. A broker that follows a post-reconciliation model will not always have to “cut back” votes in this manner, because some clients who are otherwise entitled to vote will decline to do so.
A broker that follows a pre-reconciliation model will not in the first instance invite clients to vote shares that they technically are not entitled to vote. Referring to the example above, if the client holds 100 shares of ABC Company in a margin account and 50 of those shares have been re-hypothecated, the broker will invite the client to vote only 50 shares. Mr. Smith in the above example would not be able to vote more than 50 shares in the first instance.
Isn’t the pre-reconciliation approach better because it prevents people from voting shares that they are not entitled to vote?
Not necessarily. First, it depends on what you mean by “entitled” to vote. A client who seeks to vote 100 shares—even though 50 have been re-hypothecated—really has an economic interest in 100 shares. This means that he or she will benefit from any gains, and suffer any losses, on 100 shares, not 50 shares. When the broker’s loans on the underlying shares expire, the client will hold those 100 shares both technically and substantively, so allowing that client to register 100 votes is arguably appropriate.
On the other hand: If the client above votes all 100 shares and a third party purchaser of the 50 “loaned” shares votes those shares, then 150 votes will have been offered based on the same 100 shares. So there is a legitimate debate about whether post-reconciliation is good or bad.
One other consideration: Overall, use of a pre-reconciliation model will inevitably result in lower levels of retail shareholder voting. That is because it effectively removes from the proxy voting system the types of votes described above—such that the shareholder in the example above will register only 50 votes when he or she wishes to vote all 100 votes in his or her account. An unintended consequence of pre-reconciliation is further erosion of retail voting, compounding declines that have already resulted from other factors such as the introduction of the Notice and Access model for the delivery or proxy materials and the elimination of broker discretionary voting in director elections.
Empty voting refers to a practice whereby a shareholder votes shares that he or she is legally entitled to vote in every respect, even though he or she lacks full economic interest in those shares. This can be accomplished in a number of ways, such as by a shareholder purchasing shares while simultaneously selling short all or a portion of the shareholder’s long position.
How does empty voting happen?
In contrast to over-voting, empty voting involves votes by shareholders who lack an economic interest in the company. Empty voting can occur in a number of different ways. One method would be for the shareholder to purchase the shares, but then “lay off” some or all of the risk by purchasing derivatives. Thus, a shareholder who purchases 100 shares long and then sells short the same number of shares would retain all the voting rights associated with the long shares, but have no effective economic interest in the company. Another method would be the “borrowing” of shares prior to a shareholders meeting, and the voting of those shares despite a lack of economic interest in the shares, although we have no information that this approach actually occurs in the United States.
Empty voting can be hard to define because there are numerous situations where a shareholder—particularly an institutional holder—may be both long and short the same position. One group within a large broker-dealer, for instance, may be net long a particular position, whereas another group may be net short, in each case for its own independent business reasons. Indeed, any definition of empty voting should likely have to reflect that the actor have an intent to acquire votes without incurring economic risk.
Is empty voting a systemic problem?
We are not aware of any information tending to show that empty voting occurs with any frequency. There are some reports that short sellers vote shares that they borrow in effecting short sales, but these reports are inaccurate. Even though short sellers borrow shares in order to effect a short sale, they never acquire voting rights because the borrowed shares are immediately sold to third parties in the open market. Regulations in the United States, furthermore, effectively preclude brokers from lending shares for the purpose of bestowing voting rights on the borrower.
It is possible that empty voting does occur in other countries, or in the United States where the lender is not a broker. We understand that the SEC is looking into the degree to which empty voting occurs, as well as potential new regulatory requirements.
Issuers should consider taking steps to protect themselves from empty voting, such as by imposing requirements that shareholders disclose corresponding derivative holdings in appropriate circumstances (e.g., when proposing to call a special meeting, or nominating a candidate for the board of directors).
Chicago
