# Liquidity and Control at Buffett’s Berkshire Hathaway

### 4 Responses

1. Actually, it’s not that simple.

To calculate voting power, one must – like courts in New York – use something like the Banzhaf Index of Voting Power. See, e.g., See, Banzhaf, Weighted Voting Doesn’t Work: A Mathematical Analysis, 19 Rutgers L. Rev. 317 (1965); Iannucci v. Board of Supervisors, 229 N.E. 2d 195 (1967)

The reason is that voting power does not necessarily correspond to the percentage of the votes each person can cast.

To take a trivial but illustrative example, if another shareholder had 51% of the total votes (calculated as above), Buffett’s Voting Power (with 34.9% of the votes) would be zero.

On the other hand, if another large shareholder had 49.5% of the total votes (calculated as above), and most of the other shareholders had about 1%, Buffett’s voting power would be very small, and roughly equal to everybody else’s, since 49.5% plus 34.9% of the votes is no better or more important than 49.5% plus 1% – both are enough to produce a simple majority.

PROFESSOR JOHN F. BANZHAF III
George Washington University Law School
2000 H Street, NW, Stockton 402
Washington, DC 20052, USA
(202) 994-7229 // (703) 527-8418
http://banzhaf.net/

2. Lawrence Cunningham says:

John:

You fail to distinguish between power and control. Power is the capacity to exert control. Someone entitled to cast votes has voting power even if unable thereby to dictate the outcome (i.e., control). Just because a group commanding 80% of the power can control the outcome does not mean that those with 20% have no power though they lack control.

You also assume a majority voting rule which may not be the case. On many matters these days, non-binding shareholder resolutions are put before meetings that count as success for proponents even when supported by only 20 to 30 percent of the voting power. At Berkshire, moreover, an agreement provides that, if Buffett were to command a majority of the voting power on a given matter, he would cast his votes in proportion to the voting of the other votes cast.

It is not simple at all, but quite complex.

3. With all due respect, it was not I who failed “to distinguish between power and control.” You calculated what you repeatedly called “voting power.”

For almost 50 years, the words “voting power” have been generally understood, in both legal and political science circles, to mean the ability to affect the outcome by and through voting (and only through voting).

That’s why the Banzhaf Index of Voting Power has been widely used – by me and by many others – to calculate voting power in a wide variety of situations (e.g., weighed voting, the Electoral College, recent elections in Great Britain, under the EU Constitution, etc.), why it is taught in so many colleges (including even at GWU) as well as in high schools, why many people have written computer programs to calculate Banzhaf voting power in situations where it is not apparent, and why the term appears so often on the Internet and in literature searches.

You also assert that “Just because a group commanding 80% of the power can control the outcome does not mean that those with 20% have no power though they lack control.” But I think that if, under all circumstances, someone’s vote can never affect (one way or the other) the outcome, most people would say his voting power is zero.

Strangely, exactly such a situation existed for many years in Long Island, NY. There were 6 legislators who each cast a number of votes based largely upon population. Until I did the calculations, no one realized that 3 of the 6 had no voting power – no matter how any of them voted, their votes could never affect the outcome. Not surprisingly, it resulted in a court case. To understand the underlying theory and the mathematics, see: “The Original Banzhaf Power Index Problem”
http://www2.fiu.edu/~rosentha/MGF1107/Nassau.htm

You also note that some shareholder resolutions may be counted as a success by their proponents even if they fail, often by an overwhelming number of votes. That of course is true, but one doesn’t measure “voting power”mathematically, as you purported to do, based on psychological concerns and factors, how much publicity a proposal received, whether the corporation nevertheless got the message and later made changes on its own, etc.

You are correct that voting power depends on whether the system requires for passage a simple majority vote, a 2/3 or other super majority vote, etc. That’s why the distribution of voting power has to be calculated separately depending upon the percentage of votes needed for passage – and therefore why Buffett’s voting power cannot be calculated in a vacuum according to your simple formula.

By the way, that’s also why I was able to argue successfully about 50 years ago, while I was still in law school, that “Weighted Voting Doesn’t Work: A Mathematical Analysis” [19 Rutgers L. Rev. 317 (1965)]

PROFESSOR JOHN F. BANZHAF III
George Washington University Law School
2000 H Street, NW, Stockton 402
Washington, DC 20052, USA
(202) 994-7229 // (703) 527-8418
http://banzhaf.net/

4. For anyone who may still be following this discussion of voting power, a very recent analysis/study of the Banzhaf Index can be found at “The Banzhaf Value in the Presence of Externalities”.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2393478

So, what I managed to create while only a law student is still being studied almost 50 years later!