Travis Dirks has provided a guest post on the voting power of differently sized shareholdings, which has important implications for activist investors seeking to impose their influence on a management. Travis is an expert in Nanotechnology, and received his Ph.D. from the world’s leading institution for the study of condensed matter physics, University of Illinois at Urabana-Champaign. Travis has also taught informal workshops on sustainable competitive advantage, business valuation, and the wider applications of behavioral finance and prospect theory, in addition to running a concentrated deep value/special-situations equity portfolio, which has returned 69.53% since inception in June 2006 relative to the S&P 500’s -6.08%. An entrepreneur at heart, Travis has co-founded one successful local business and one technology startup. He is currently working on a book – Voting Power in Business. Travis can be reached at TravisDirks@gmail.com.
(I would greatly appreciate feedback on whether the information below is new and/or useful to you)
Process over outcome. The watch words of all great investors. By thinking probabilistically in terms of relevant long-term averages, such investors gain control over a field swayed by random events. Yet, it appears that this method of thought has not yet been thoroughly applied to shareholders’ only means of control: the shareholder vote. When probabilistic thinking is applied to the proxy battle the activist investor, and those of us who rely on him to catalyze our deep value investments, gain a counterintuitive and valuable edge! Here, I outline the strategic value of voting power – a measure of a shareholder’s true influence – via two examples: a simplified hypothetical and a real world company.
How does ownership differ from power?
Pop quiz: 1) Does 10% ownership of a company imply 10% influence on the vote? 2) Does buying 1% more of the company imply proportionally more power? 3) Is your influence on the vote impervious to other shareholder’s buying or selling?
Answers: No, no and NO!
Voting power is not related to ownership in a straightforward way – it is a function of the entire ownership structure of the company, i.e., how much everyone else owns. Voting power measures the percentage of all outcomes where a voter gets to decide the result of the vote. In fact, you can gain more or less power depending on who you buy from! The counterintuitive nature of these answers hints at an opportunity for the enterprising investor.
A simple example of voting power analysis
Alice and Bob each owned 50% of a successful rapidly growing small business. Like so many before them, they lost sight of cash flow and needed money fast. With no bank to turn to, they decided to sell 2% of their business to uncle Charlie. So the ownership structure looks like this: Alice –> 49%, Bob –> 49%, Charlie –> 2%. Does Charlie, with only 2% ownership, really have any influence on the company? The astounding fact is that Charlie now has just as much power over the business as Alice and Bob. When matters come to a vote, Alice, Bob, and Charlie are in exactly the same situation: to win the vote they each have to convince one of the other two to agree with them and it doesn’t matter which one.
Not everyone is as fortunate as Charlie – a voter can have drastically less control than his ownership would indicate. In fact, he may even have no control at all!
Let’s rejoin the story a few years later. Charlie now owns a third of the company, as do Alice and Bob. They’d like to raise more capital to buy out a weak competitor so, at Charlie’s suggestion, they do something clever: they sell Dan 20% of the business (new ownership structure: Alice –> 26.66%, Bob –> 26.66%, Charlie –> 26.66%, Dan –> 20%). Interestingly, though Dan paid for 20% of the business, he received zero voting power! The reason is that there is no possible voting outcome in which Dan could change the result by changing his vote. All possible voting outcomes have a winning majority even without Dan’s vote! Effectively, Dan is wasted conference space. Thus:
Rule 1 of voting power analysis: a voter can have drastically more or less voting power than his ownership would indicate.
Driving strategy via voting power analysis: Breitburn Energy Partners
What strategically valuable information can such statistical analysis reveal for a large public company? Consider such a company from my (and Seth Klarman’s) portfolio: BreitBurn Energy Partners L.P. (NYSE: BBEP)
BBEP’s is a beautifully intricate story in which the oil crash, the market crash, an angry majority shareholder, a convenient bank loan covenant, 5 years of hedged production, and the fleeing of dividend-loving stock holders combined to create the easiest purchasing decision I’ve ever made! Voting power analysis sheds new light on one part of this story – a dispute between the majority shareholder and the company over voting rights.
In June 2008 the board decided to give the limited partners the right to vote on who would sit on the board of directors—with one caveat. No one share holder could vote more than 20% of the company’s shares. In the event that a shareholder owned more than 20% of the outstanding shares, the final vote would be counted as if the rest of the shareholder’s votes did not exist. This understandably upset Quicksilver Resources Inc., who held 40% of the outstanding shares. Strategically, how should Quicksilver and The Baupost Group (the other large shareholder) react to these unique voting rules?[i]
Figure 1: Percentage ownership (blue), and voting power under two schemes - standard (red) and BBEP’s 20% cap (green) for Quicksilver Resources Inc. and The Baupost Group, the two largest stakeholders of Breitburn Energy Partners.
In Figure 1 the blue bars show Quicksilver’s and Baupost’s percentage ownership of BBEP as of March 2010. The red bars show each company’s voting power under a standard voting scheme. Notice that under a standard voting scheme Quicksilver has drastically more voting power than their ownership indicates and Baupost has drastically less. This disparity, with some shareholders having more power and some less, is closer to the rule, than the exception. Under the 20% cap rule (green bars), however, Quicksilver’s voting power is cut in half and Baupost’s tripled, but both have influence that is more in proportion with their ownership – a powerful insight that could have been used in BBEP’s defense in the inevitable lawsuit that followed.
The lawsuit was settled and a voting system in which Quicksilver got to vote all its shares (but others who crossed the 20% threshold did not) was agreed on. The resulting power distribution is now exactly the same as under a standard voting scheme (red bars). Therefore, in the event of a disagreement between Quicksilver and Baupost, Baupost’s chances of success are much worse than their ownership percentage would suggest.
How to use voting power analysis to minimize influence loss?
The settlement also seemed to indicate that Quicksilver would be selling down its majority position. It has already sold nearly a quarter of its holdings, most of which went to a new share holder, M.R.Y. Oil Co. Now, assuming Quicksilver would like to retain some ownership, what strategic insight can voting power analysis lead to? Because voting power is a function of both individual ownership and the overall ownership structure, it is actually possible to minimize your lost voting power (on a per share basis) by strategically selecting low-impact buyers.[ii]
Consider two options open to Quicksilver to drop below the 20% ownership threshold: selling a third of its holdings (10% of the company) to the second largest stakeholder, Baupost, or selling it to small shareholders on the open market.[iii]
Figure 2: The current voting power structure at Breitburn Energy Partners (blue: ownership, red: voting power), as of July 2010, and two hypotheticals in which Quicksilver Resource Inc. sells another 10% of the company. In the first scenario, the shares are distributed among the smallest shareholders (orange bars). In the second, the shares are sold in a lump sum directly to The Baupost Group (pink bars). The inset shows the percentage change in power from the current situation for each hypothetical.
Figure 2 shows each company’s ownership stake (in blue), as of July 2010, along with the associated voting power (in red). First, note that Baupost’s voting power has improved significantly from March (see Figure 1) without buying any shares because the ownership structure has changed. Second, Quicksilver’s voting power changes depending on whether it sells its shares to the smallest shareholders (in orange) or to Baupost (in pink). When Quicksilver sells to the masses, Baupost’s power again increases substantially, even though they have not increased their holdings by a single share! (The same is true for M.R.Y. Oil Co.) Thus:
Rule 2 of voting power analysis: voting power can be increased by influencing others to buy or sell!
In the inset in Figure 2 we can see that if Quicksilver sells its shares to Baupost, Quicksilver loses nearly 50% of their voting power. Whereas, if they sell to smaller shareholders they lose only 37% of their voting power. Thus, even a simple application of voting power analysis has profound implications – given the choice, Quicksilver can buffer its loss of influence by a whopping 13% by distributing its shares to smaller shareholders, rather than selling a lump sum to Baupost.
Voting power analysis: what else is it good for?
Stakeholders can use voting power (and related) analysis to help guide key strategic decisions like:
- Who to buy from and who not to buy from to maximize their purchased control
- Who to sell to and who not to sell to minimize any lost influence
- When to be greedy by identifying situations where the purchase of a few more shares will result in a large increase in influence
- When not to be greedy by identifying situations when the sale of just a few shares will result in a large decrease in influence
- How to increase their influence without buying any more ownership
- How much of a company can be sold, while giving up ZERO voting power
In conclusion, control isn’t worth much – until it is. When trying to change the course of a business in crisis, voting power analysis might prove essential. As voters form coalitions, the effective sizes of the voting blocs grow and so can the disparity between voting power and percentage ownership. It is here in the crucible where, I believe, voting power analysis plus intelligence about key voters’ stances can provide a real edge to investors trying to decide whether the proxy battle is worth the expense and, if so, how best to win it.
Again, I would really appreciate feedback – especially on whether such voting power analyses are commonly applied either in the boardroom or by activist investors.
Travis Dirks, Ph.D. (TravisDirks@gmail.com)
I’d like to thank my co-authors, Radhika Rangarajan and Guy Tal for their insightful conversations that both inspired and fleshed out these ideas as well as for their heroic efforts in helping me to (hopefully) make this tricky subject understandable!
[i] I have made two main assumptions for ease of calculation. The first is that all shareholders vote their shares. The second is that the shares held by groups and individuals with small enough position to not file a 13d are held in equally sized small pieces. While a more realistic assumption, such as a power law distribution in shareholder size, will change the values of voting power, it will not change the main results: percentage ownership alone does not equal voting power, allowing for full optimization of voting power. Also, a Shapley-Shubik index is used to measure voting power.
[ii] It is also possible to optimize in other ways. For instance we can buy shares to maximize our voting power gain or even our voting power gain plus competitors’ voting power loss.
[iii] Here I’ve assumed that Baupost would get to vote all 25% of its shares, as it would in most other companies, after its own lawsuit. Interestingly, if Baupost were to suffer the 20% cap, selling to Baupost or the masses becomes nearly equivalent options for Quicksilver (Q = 28% , B = 22% , M = 7% ).