The Value of a CEO

Steven Ballmer announced today that he would retire. Microsoft stock shot up immediately by ~$2.18 or 6-7%. Given 8.33 billion shares outstanding that’s an increase in value of about $18 billion dollars. Of course that’s embarrassing for Ballmer but the lesson cuts both ways. If Ballmer’s exit and replacement with an unknown is worth $18 billion then hiring the right CEO at $27 million annually, the average annual pay for the 100 highest paid CEOs in America, looks like a bargain. Small differences are a big deal for large corporations, you know like a marginal… something or other.

Ballmer Retirement

Hat tip: Justin Wolfers.


Talk about a golden parachute!

I'm not remotely qualified to be the CEO of a large corporation. But then neither are most of them.

Actually, it now is apparent that an important function of the CEO is to figure out the optimal time to retire, the time that maximizes shareholder value. Ballmer has created $18B of shareholder value through his decision, more than earning his golden parachute. :) :) :)

I'm a bit surprised we don't see CEOs being assassinated by shareholders. The Mafia will kill someone over a paltry hundred thousand dollars. When tens of billions are at stake, an individual large shareholder might be losing hundreds of millions by having a bad CEO in place. One rifle bullet could change that in an enormously positive way. There's even a humanitarian aspect to this -- think of all the jobs which are created when a company goes from being sick to being prosperous. It's immoral to let one man (or woman) stand in the way of that. Where are the good assassins?

Please set the whole story in a not so distant future (2030) and I'll buy the book, even go to the movies for the adaptation.

Assassin's Creed: Fiduciary Out.

I can see it now, the greatest game of my generation, if Ubisoft will stop dicking around with DRM that ruins games.

Cold Cash War, Robert Asprin. Could probably do with an update though. :-)

This sounds like a combination of Robin Hanson's "Fire the CEO market" and Jim Bell's (much more controversial!) "assassination politics".

But shoot the head, so we can harvest the useful organs for the needy.

CEOs have useful organs? Who knew?

"I'd recommend harvesting his organs, but those probably won't work either." - Phil, Prince of Insufficient Light.

I think it's true that a bullet to a CEO's head could be worth millions of dollars, but, unlike the mafia, violent characters aren't attracted to the business of buying and selling stocks because willingness to commit violence isn't a beneficial character trait to have in investing. Violent, ruthless individuals are attracted to the mafia because it is a desirable trait to have when pursuing illicit business activities that can result in arrest and/or death. That would be why CEO murders aren't a popular occurrence.

That means the market is not at equilibrium. Capital is being held captive in the boardrooms. We need market innovators that remove these barriers against shareholders receiving the full value of their investments.

Tyler, you must be a blast at parties ;)

If you are a such a large shareholder that the replacement of the CEO would be worth millions to you, you would be controlling enough shares that you should be able to influence the decision to merely fire him.

Most corporate boards think that their management team is above average, and deserves an above average compensation package.

Similarly, voters have a low opinion of Congress but a higher opinion of their congressman. Parents may have a low opinion of their school system but generally have a higher opinion of their neighborhood school. It's a cognitive bias, perhaps based in the functional in that we have to think our own children are worth nurturing.

This seems to imply that Balmer has been paid far too much for the last few years. Perhaps it merely represents the extreme uncertainty (low information content) of the relationship between pay and value for CEOs.

OTM. If anything this shows that boards are very bad at predicting CEO performance, so they should probably not waste company money by paying top dollar for a superstar CEO.

The post-retirement bounce in and of itself implies nothing about Ballmer's past performance. It implies, as Alex suggests, that going *forward* the market *expects* that a generic, unnamed CEO will create $18B more value than Ballmer would have. One could argue that the large bounce suggests market confidence in the ability of the board to pick a successor.

To evaluate Ballmer's performance, one needs to evaluate MSFT performance over the entirety of his tenure, including this post-retirement bounce since every CEO can choose to leave at any time and, eventually, always does leave. For example, if MSFT performance prior to the bounce was poor but performance including the bounce was acceptable, that would imply that any loss of shareholder value caused by a CEO could easily be recouped when the CEO retired. I have no idea whether MSFT has performed well in recent years.

Alex is correct that any large stock price movement, either positive or negative, in response to the hiring or resignation of a CEO suggests that the market perceives, rightly or wrongly, that the choice of CEO is important to shareholder value. Muted stock price movement would suggest either that the change was largely anticipated or that market participants are largely indifferent to the choice of CEO.

Seems to me that Steven Ballmer was way overpaid, like most CEOs. Though obviously Alex believes that only if Microsoft had been offering a billion dollar annual salary some rockstar CEO that had previously found XXX million below his reservation wage would have come forward and they wouldn't have given Ballmer the job.

Logically, that is probably not be the case, because market participants value Mystery CEO X.Y., representative of "most CEOs", much more than Ballmer.

What is the logic of this statement??????

"If Ballmer’s exit and replacement with an unknown is worth $18 billion then hiring the right CEO at $27 million annually, the average annual pay for the 100 highest paid CEOs in America, looks like a bargain." ?????

If someone poorly manages a firm,


If Alfred E. Newman, or me,

Does Better, Notwithstanding the incompetency of the predecessor,

I should, or Alfred Should, get the difference???

Hmmm. Have you ever thought that if the CEO announces a retirement the year before he retires, and the stock goes up, perhaps, instead of a retirement bonus, there should be a pay reduction in the last year, or a clawback?

I bet instead that Microsoft will be broken into different lines of business. Easier to do if the CEO is leaving. In other words, the valuation may be including a spin off of pieces as well.

Or, maybe it is entirely the potential of a new CEO.

Or, maybe the foreign bribery allegations are getting a little too hot and close to home and its time to leave.

Or maybe the share prices will go down tomorrow.

The logic starts by recognizing that not every CEO can be below average for the same reason that not every CEO can be above average. So, if Ballmer is worth $18B less than a generic, unnamed CEO, it suggests, at the very least, that there should be massive inequalities in CEO pay: the generic CEO should make a lot more than Ballmer would have had he stayed on as CEO going forward. (Yet, one more illustration that meritocracy implies income inequality.)

There is a zero lower bound on nominal income (for any employee, not just CEOs) for much the same reason that there is a zero lower bound on nominal interest rates. Workers always have the option of not taking a job that pays a negative income. Also, if a firm is forced, for whatever reason, to hire a negative marginal product CEO, the firm could always turn that CEO into a zero marginal product employee by putting him/her into the ZMP room described by Tyler in earlier posts. Thus, of the 7 billion people on the planet that could (nominally at least) be chosen to be the next CEO, the worst one should be paid at least zero. Since there should be massive income inequality among potential CEOs, that suggests that the best choices, should they be selected by the board, would indeed merit very high compensation, as Alex suggests.

Note, nothing in this argument necessarily implies that the *current* 100 highest paid CEOs, nor Ballmer's eventual replacement, are "a bargain", just that there exist 100 people such that if they were CEOs, they would be worth more than $27M annually. Thus, while claims along the lines that "no one should make as much as CEOs do" are largely off the mark, one could legitimately claim that a *different* group of 100 people should be the ones making these insanely high levels of compensation. In short, we should argue about *who* should be in the top 1%, not *whether* there should be a 1%.

For context, you should also show the Thursday close/Friday open for other companies whose stock prices tend to covary with Microsoft's, so I can be sure that nothing else happened overnight on Thursday.

". If Ballmer’s exit and replacement with an unknown is worth $18 billion "

Assuming an instantaneous change in nominal market capitalization is identical to the change in the actual "worth" of a company.

Instantaneous market capitalization of a firm is not a directly realizable asset, is it?

I generally have the view that CEO's are not paid too much. However, I do not think that this evidence proves the point. When the value of Microsoft stock goes up by $18 billion after Balmer announces his retirement, this suggests that Balmer was being paid too much (he should have been paying Microsoft $17 billion to keep his job). A downward movement in the price of Microsoft shares would have been a more convincing example.

+1 to wd40. Some CEOs destroy value.
Also, the original post assumes Microsoft was absolutely perfectly and correctly valued both before and after the announcement. I'd think we've been through enough bubbles to see through that.

Returning just to add that a free Windows phone is overpriced. We've been trying for three weeks to download free apps. Still trying.

What we do not know is how good we are, a priori, of judging a CEO's performance, and how easy it is to replace one.

Imagine that, out of a pool of 100 CEOs, 2 will triple market capitalization after 3 years in the business, and most just manage to perform along industry standards. Offering a lot more money for the position would be worth it if I manage to get one of the two amazing CEOs, but if I just can't separate them from the pack, doing an expected value calculation for the entire population and paying them all the average just doesn't make sense, because the extra pay is not really helping you find the good executive.

Offering a higher salary entices more applicants, but if you can't tell the good from the bad, and you can't guarantee that your extra applicants will have better quality than at a lower salary, offering more money up front will do no good. Thus, if you want to get a great candidate, you either have to pick a known value, and pay him enough he'll jump ship, or have an incentive based system that has a good projection of how an average CEO would perform.

HAHA! I see what you did there...

Apple's stock price went up, not down, when Steve Jobs stepped down:

Maybe we can avoid being a little too simplistic about CEO value? (Or maybe Jobs and Ballmer were equally valuable.)

Nobody was pricing in a significant chance that Ballmer would die without there being a orderly transition of power. In the case of Jobs there was the risk of Steve dying from pancreatic cancer with no orderly transition plan in place and a bunch of turf wars and infighting to occur in the power vaccum.

I doubt that CEO pay has much to do with CEO value. I think it has a lot to do with other forces. When the CEO is paid more, people just below him can be paid more, and so on. In this case, the price change reflects Microsoft's poor performance recently. I know people who work at Microsoft. They have an unusually dysfunctional culture. (The Surface is a stunningly bad product.) That might be one reason Microsoft has done poorly. Another is the Christensen Innovator's Dilemma idea, that leading companies become smug. I believe Microsoft investors are too sure what has caused the decline -- which will continue.

Sure huh? Are you trying to tell us that you are now shorting MSFT?

Buy SP500 index funds, enjoy the show from the cheap seats. MS has too many internal and external problems for a return to rock star status. Perhaps they will find modest success with a midlife career change, but who knows ...

The hope is the next CEO will be a financial type who will slash R&D, increase debt, and milk Windows and Office for maximum profit. Instead of trying to play the hero.

Slash R&D?!?!?! But that's the only thing of value coming out of Redmond these days. The value may not be realized by Microsoft, but what a waste that would be for the rest of us. Spin off the scientists and let them license the amazing tech. Surface could have made billions in the right hands.

The hope is the next CEO will be a financial type who will slash R&D, increase debt, and milk the declining monopolies. Instead of trying to play the hero.

Isn't part of problem the "heads I win tails you lose" asymmetry in possible outcomes for CEOs? Sure ex-post we can have some (still not convincing) idea of what a CEO was worth to the company, but it's really hard to tell in the moment or ex ante. Thought experiment - say the probability distribution of some CEOs value to the company ranges from -$1bn to 1bn (2 standard deviations). Then in theory his pay should be 0 right? But in reality his pay will be a few million. If the company then does well, he'll get paid many millions more. If it does poorly, at some point he'll be fired. But that means his worst case outcome is the median of the "real" distribution. This ignores the possible efficacy of options, but I think the point is still valid?

This is a good way to look at it. The view of, "see how much the CEO impacts the stock price - he should be paid hundreds of millions!" is just absurd and overlooks the obvious uncertainty in the situation.

The question is: if it's worth $18 billion to kick him out, why didn't the shareholders do it before? The apathy and powerlessness of shareholders of public companies is one of those things that I REALLY don't get. Any answer seems to require extremely unrealistic assumptions.

On the other hand, when Steve Jobs Announced his retirement, AAPL dropped 5.13% (21$/share). that made his worth about 19 billion.

Steve Jobs is the anti-Steve Ballmer.

My mental meandering on this issue: to what extent does the jump in share price reflect any real changes in Microsoft's prospects because of a CEO departure vs. just irrational response/noise? But then I also find it difficult to second guess the many shareholders and new investors who are putting their money on the line. Even if the market isn't perfectly efficient, it is efficient most of the time.

CEOs are so important because they have the potential to do so much harm. We pay millions of dollars per murderer to keep murderers off the streets too. So both murderers and CEOs are important to society because they have the tendency to disproportionately affect the lives of others more than most. That said, reducing the costs to companies and societies of murderers and CEOs might be worthwhile.

Isn't the more likely explanation that the stock market overreacts to news, both good and bad? If it were so obvious that Ballmer was holding the company back that much, the board could have fired him a long time ago. My prediction is that MSFT will drift back to a more reasonable valuation over the next few weeks or months as the news bump fades.

Better go short sell! You could make some good money!

Microsoft could have easily attracted someone capable of destroying just as much value for about $50k a year.

The CEOs number one job is capital allocation, and Ballmer couldn't have done a worse job. If the next CEO does nothing more than cut the r&d budget in half and use the cash to pay a bigger dividend, stock goes up by $10. So you're right, any outsider can see that Ballmer is using the Microsoft cash flow to have fun playing with lots of toys, like Bing, MSN, and X-box.

Yes, and Obama should be paid $136 billion dollars by publicly traded companies, because that is how much the stock market tanked when the fake tweet about him being injured went out.

So if Ballmer was assassinated in a massive terrorist attack while being protected by the secret service and Microsoft stock went down, it would clearly be because Ballmer was so valuable? Even if Ballmer was not CEO but merely a member of the board?

Steve Ballmer owns 333 million shares so the jump in value gave him another $725 million not a bad "golden watch" retirement gift.

I think your logical leap went too far. All the market said was that Balmer NOT being there had the potential to make Microsoft CEO actually appreciate in value. CEO pay somehow being linked to stock price is a ridiculous idea. If they announced a squirrel as the next CEO and the stock dropped by 50% would that mean Balmer was underpaid by 50% of the stock price?

Balmer's clearly one of the worst CEOs in history whose greatest achievement seems to be not setting his money printing press on fire.

"Balmer’s clearly one of the worst CEOs in history whose greatest achievement seems to be not setting his money printing press on fire."

If this make him one of the worst, you must define "worst" pretty broadly.

So... Alex's logical leap went too far, but you are going with a squirrel analogy?

Alex: "If Ballmer’s exit and replacement with an unknown is worth $18 billion then hiring the right CEO at $27 million annually, the average annual pay for the 100 highest paid CEOs in America, looks like a bargain. Small differences are a big deal for large corporations, you know like a marginal… something or other."

That assumes that boards are picking the right CEO, rather than a crony.

I don't think this proves that we should pay good CEO's to come on board and stay, it proves that we should pay bad but entrenched CEO's to leave.

Well, "prove" may be a little strong.

Pay among big company CEO's is a status symbol, and does little to motivate behavior. The marginal utility of an extra dollar declines to close to zero at the level of pay most of these people receive. They're motivated by many things, but at that level of wealth, money doesn't really do much for them except notch their belt. And, as any good economist would tell you, value creation should be the upper limit of pay, and should not the benchmark for pay. Instead, things like the supply of able replacements play into it. In the real world, power politics and leverage points heavily influence remuneration.

“If Ballmer’s exit and replacement with an unknown is worth $18 billion then hiring the right CEO at $27 million annually, the average annual pay for the 100 highest paid CEOs in America, looks like a bargain."

1) breathtakingly spurious, superficial logic

2) Arch Defender of the One Percent!

It's spurious to say that $27 Million is not much to pay for someone who can create billions of dollars of value?

Alex is the "arch" defender of doctors and lawyers?

When Fiorina was canned from HP, the stock was up 10.5% at one point that day:

Mother Jones cites your post, Alex:

Surprisingly (NOT) they neglected the lesson you draw from this episode.

Perhaps CEO pay is a kind of option value - they are not paid the value they add, but the option value of potential added value above the next candidate. The degree of uncertainty as to the ultimate value add is very high with multiple paths and a lot of randomness and noise. Add to that the CEO himself faces optionality along with the significant agency costs. Perhaps that is why CEO pay is so complicated, with options, RSU's, contingent payouts, golden parachutes, etc. Sounds like a good paper for an hungry academic.

Consider the alternative hypothesis:

a) CEOs have a large impact on the value of their firm (could be a net positive or net negative).

b) CEOs may have a large impact on value, but sometimes it's negative, sometimes it's positive, and on average it is zero or small. Therefore, since their value added is small or zero, they are overpaid.

Isn't the evidence shown in Alex's post consistent with this hypothesis? That is, Ballmer has a negative effect, so his removal and replacement with someone with a "less negative" effect (or even a marginally positive one) boosts the stock price.This is true despite that fact that Ballmer and his successor can both be overpaid.

You don't think it's worth $27 million to reduce the chances of losing $18 billion?

This assumes that additional CEO pay leads to hiring better CEOs. Do we have any reason to belief this? It seems that some basic price = quality assumption is at work here when there are all sorts of reasons why that might not be the case.

Maybe I wasn't clear. Under the alternative hypothesis, CEOs do not add value on average. So in aggregate, they would not reduce the chances of losing firms money, regardless of how much money you throw at them.

My point is that this hypothesis is also consistent with the evidence that Alex presented, but it precisely the opposite of Alex's conclusion.

What I was trying to illustrate is not that the alternative hypothesis is necessarily correct, but rather that the evidence doesn't really support either hypothesis.

My trouble with your hypothesis is it assumes that the CEO is optional. Like R&D, a firm can have a lot of R&D, or no R&D, you can work out the average benefit of having R&D. But any standard firm is going to need a CEO or something equivalent so I'm not sure it's meaningful to say a CEO adds or detracts value on average.

Working out the relationship between dollars spent on the CEO and firm performance, that might be more meaningful.

No one would be a better CEO than Steven Ballmer.

There's got to be some 'pure beauty contest' effects going on. Balmer was bad, but he wasn't $18b bad - however, many people see him as such, because scope insensitivity causes us to think in terms of "ok; bad; really bad; steve balmer bad" with no real cardinality.

Alex doesn't seem to understand thinking at the margin. Let's say I'm the office IT guy. I could subtract millions from the value of my company by screwing up and deleting some files. But just because I have the potential to shift the company's bottom line significantly does not mean my salary should be correspondingly high. There are plenty of people out there with the skill set to make them competent IT guys who won't screw up badly. Because there are so many workers with the relevant IT skill set and my marginal value over many of them is relatively low, my salary is likely to be pretty average.

Value over replacement CEO is the relevant margin.

Any chance the value can be attributed to a well-functioning succession plan and signaling generally good corporate governance?

(In case you chuckled ... Hey! That wasn't meant to be a funny question ...)

I'm surprised no one has pointed out that -$18B is a very shaky measure of value given that very few of the 8.33B shares of Microsoft are in play at any time.

These analyses always underscore to me how poor our measures of wealth are. I guess it's better than nothing, but it's also a completely notional value and should be treated as such. Relatedly, since most wealth is notional and contingent, redistribution is that much worse than it is on its face.

Is it really a true that hiring the right CEO is worth $27 million? Isn't it really that hiring a bad CEO simply destroys billions of dollars?

I know there have been studies done that show that most CEOs do not impart any additional shareholder value than another person with similar experience and credentials. Obviously, that level of expertise is worth something, but most CEOs seem to be overpaid. Very few create longterm shareholder value above what is expected.

The high price of CEOs is a matter of corrupt boardroom cultures, an inefficient tax system, and legal obstacles to shareholder activism.

The bigger problem for Mircosoft is not Balmer, but their business model. Microsoft became huge not by being an innovator, but offering the benefits of industrywide standardization which more than compensated for their inferior software copied from those created by its smaller, but more innovative rivals. The market's computing needs are now going into another direction - one which Microsoft is not well prepared for.

Comments for this post are closed