1, 2 or 3 Years? How Should Boards Decide The Frequency Of Say-On-Pay Votes?
Monday, September 29, 2014

Although the initial ruckus over how to determine the outcome of executive compensation say-on-frequency votes has subsided, I still maintain that the Securities and Exchange Commission botched the rule (§240.14a-21(b)).  More importantly, it seems many boards don’t understand the challenges with ascertaining and applying stockholder preferences after the votes have been counted.  At the time, I advocated for the use of a voting system devised by an 18th century French mathematician, Charles de Borda.  Monsieur Borda observed that when voters are presented with three or more choices, the choice that receives the most first votes in a plurality voting system may not the choice with highest overall standing.  His solution was to allow voters to rank preferences.

Marie-Jean-Antoine-Nicolas de Caritat, the Marquis de Condorcet, was also concerned with the problem of vote aggregation methods.  A contemporary of Monsieur Borda, the Marquis proposed pairing of alternatives in a round robin of imaginary votes.  Let’s imagine that a three member board of directors meets to consider the frequency of say-on-pay votes.  The directors’ preferences are in the following order (> indicates that the left hand valued is preferred over the right hand value):

Votes

Order of Preference

A

1 yr > 2 yrs > 3 yrs

B

2 yrs >3 yrs > 1 yr

Γ

2 yrs > 1 yr > 3 yrs

There are three possible one-on-one match-ups – one year versus two years; two years against three years; and one year against three years.  In a vote of one year against two years, two years wins because both directors B and Γ favor two years over one.  In a vote of two years against three years, two years is favored by all three directors over three years.  In the match-up of one year against three years, one year is the winner because it is favored by directors A and Γ.  In this example, we can conclude that two years is the preferred choice of the three directors because it prevails in every match-up.  Two years is the Condorcet winner in this example.

There is an interesting problem with this approach, however, that illustrates how social preference doesn’t always perform as one might expect.  Assume the following preference order:

Director

Order of Preference

A

1 yr > 2 yrs > 3 yrs

B

2 yrs > 3 yrs > 1 yr

Γ

3 yrs > 1yr > 2 yrs

Thus, we find that a majority of the directors prefer 1 year over 2 years; 2 years over 3 years; and 3 years over 1 year.  In this case there is no Condorcet winner because no frequency prevails in every match-up as it did in the preceding example.  What makes this example even more odd is that one might expect that one year would be preferred to 3 years because 1 year is preferred over 2 years and 2 years is preferred over 3 years.  The expectation is not met because while each director’s preference is transitive, the preferences of the entire board is intransitive.

 

NLR Logo

We collaborate with the world's leading lawyers to deliver news tailored for you. Sign Up to receive our free e-Newsbulletins

 

Sign Up for e-NewsBulletins