Cumulative Voting for Directors

Bill Sjostrom —  19 January 2006

While working on my last post, I discovered that Hewlett-Packard’s certificate of incorporation provides for cumulative voting in the election of directors. This made me curious as to how many other public companies have cumulative voting so I googled it. I came up with this article which says about 10% of the companies in the S&P 500 have cumulative voting.

Cumulative voting has absolutely no impact on an uncontested election. It could, however, impact a contested election because it reduces the minimum number of votes an insurgent must secure to get at least one director on the board. But as argued strenuously by opponents of the SEC’s failed shareholder access proposal, i.e., big business, a board with insurgents is likely to become politicized and balkanized to the detriment of shareholders. Why then would companies like HP provide for cumulative voting (assuming they are not forced to through a Rule 14a-8 proposal)? My perhaps cynical guess is that they view the very low risk of a contested election being impacted by cumulative voting (according to this paper during the period of 1996-2004 there were on average only two contested elections per year at companies with market caps in excess of $200 million) as outweighed by the public/investor relations benefit of having cumulative voting. As HP recently stressed to the SEC, cumulative voting “is viewed as stockholder-friendly� and “gives stockholders a meaningful role in the election of directors.� Also note that the influential Institutional Shareholder Services (ISS) generally supports cumulative voting (see here).

3 responses to Cumulative Voting for Directors

  1. 

    Bill,

    I realize you have a lot of practice experience in this area. Wouldn’t it make sense, from a minority interest or founder perspective, to have both a shareholder agreement and cumulative voting at time zero? The shareholder agreement obviously goes away when the company goes public, but there there could be some robust supermajority provision that makes it highly likely that cumulative voting will remain. That might be why relatives of the founders were well-situated to wage a proxy fight when HP attempted to merge with Compaq. Just a thought. bh.

  2. 

    I agree it would be interesting to know how these companies end up with cumulative voting. As for HP (or any public company), I would be surprised if it is a remnant from its founding. There are much better ways to deal with minority shareholder concerns at founding than cumulative voting, i.e., a shareholders agreement. Also, companies routinely “clean-up� their certificates before going public which includes removing things like cumulative voting. Maybe I’ll have a research assistant look into this.

  3. 

    Bill, this is an interesting post. I realize your are characterizing the opponents of the failed SEC shareholder access proposal when you note the presence of directors who represent minority interests will lead to “politicized and balkanized [Boards that operate] to the detriment of shareholders.”

    That argument, of course, is absurd. It just means that instead of pro forma board meetings where directors rubberstamp the recommendations of management, there might actually be someone in the room who says, “Hey, that is an agency cost.” In turn, management will actually have to defend their overreaching or lassitude. As I tell my students, Socrates walked the streets of Athens asking uncomfortable questions, and they made him drink hemlock.

    Regarding HP, it is possible that the cumulative voting provision is a carryover from the founding of the company. (If I were a minority investor, I would ask for cumulative voting just so I could be in the room when the majority voted to steal from me.) It may still be there because of something in the charter or certificate of incorporation. Or maybe HP just sees some benefit on being on the ISS good guy list. I would be interested to know if cumulative voting typically flow from a predictable sequence of path dependent events. The 10% figure suggests a decent and tractable sample size.