Academic

American public corporations are going to have to institute Say on Pay 

advisory votes as a result of a provision in the Dodd-Frank Act. Say on Pay 

has been required in the UK since 2002. The UK experience suggests that 

insofar as Say on Pay has had any sort of limited effect, it has only been 

for poorly-performing companies. Most debate over Say on Pay so far has 

been a battle within the general war over shareholder rights, with 

pro-shareholder types supporting it and pro-management types opposing it.

In "Say on Pay's Bundling Problems," Andrew Lund tries to change the focus. 

He is dissatisfied with the limited impact of the British (and now the 

American) version of Say on Pay, and he is looking for a way to increase 

that impact. He argues that the current version has a bundling problem: if 

shareholders unhappy with a company's compensation scheme vote "no," the 

CEO is likely to take that as a personal insult. The result may be lowered 

performance or the CEO's departure. Shareholders are willing to pay that 

cost only at poorly-performing companies, even though they may prefer 

different pay packages at well-run companies too. Lund suggests an "ex 

ante" version of Say on Pay to address this problem. Shareholders would 

vote on the outlines of compensation schemes to be offered to the next CEO, 

before she is chosen. Thus, there is no personal insult, and shareholders 

can vote purely on the basis of their preference on pay systems.

This is a clever idea, well worth considering. I would assume that 

opponents of shareholder power will oppose this version of Say on Pay as 

well. Indeed, perhaps they will oppose it more strongly--as Lund 

recognizes, insofar as his scheme allows shareholders to flex their muscles 

more easily, it is for that reason objectionable to those who want to keep 

them bound. I do wander whether the CEO offense problem within the current 

scheme is really as severe as Lund posits. Are CEOs really that 

thin-skinned, so that a non-binding vote against their pay package will 

send them off in a huff? Would one really want to retain such a 

thin-skinned narcissist? But if indeed enough shareholders do believe their 

CEOs are that thin-skinned and yet do want to keep them happy, then the 

problem he identifies is real (for those who want Say on Pay to be an 

effective shareholder tool).

Also, it seems to me that the pattern observed in the UK of shareholders 

voting no only at poorly-performing companies is not really all that bad. 

Doesn't that fact make the system in effect more performance-based? Lund 

marshals evidence that this is not what institutional investors and proxy 

advisors say they way to do, but even if it is an unintended effect, it 

seems to me to have some redeeming features.

One concern about Lund's scheme. What compensation scheme is best may 

depend rather heavily on the particular CEO. Is she an insider or outsider? 

Experienced and nearing retirement or at an earlier career stage? Is she a 

well-known star, or an obscure person in whom the company sees promise? 

What kind of pay does she receive at her previous job? And so on. Voting on 

the compensation scheme in advance of knowing the answer to these questions 

may be difficult and misleading.

Lund hints at a multiple equilibrium possibility. The insulting the CEO 

effect would diminish or disappear if shareholder votes against bad pay 

schemes were more common outside of poorly-performing firms. But given the 

current pattern, a no vote does have that message, and hence shareholders 

will vote no only where they are unhappy with the CEO. If this is right, is 

there any way to encourage the widespread no vote equilibrium as we start 

the Say on Pay system here? If in the first year activists and proxy 

companies focused on no campaigns in a group of prominent companies with 

good CEOs and strong performance but dubious pay schemes, might that keep 

us from going down the road that Lund fears?

I would like to see Lund expand upon and defend some important details. 

Should all public corporations be required to offer his version of Say on 

Pay, or should shareholders be able to opt out? Or might an opt in version 

be better? Why should the shareholder vote be advisory rather than binding? 

Should a company use Lund's scheme along with more traditional Say on Pay? 

Lund suggests the latter possibility in note 202, but does not elaborate. 

The question is important, since for now the traditional Say on Pay is the 

law for the U.S., and given its newness, I doubt that will change in the 

immediate future.