Are CEOS paid enough? A look at sudden deaths

From Bang Dang Nguyen and Kaspar Meisner Nielsen:

An efficient managerial labor market should compensate executives according to their contribution to shareholder value. We provide novel empirical evidence about the relationship between executive pay and managerial contribution to value by exploiting the exogenous variation resulting from stock price reactions to sudden deaths. We find, first, that the managerial labor market is characterized by positive sorting: managers with high contributions to value obtain higher pay. We find, second, that executives appear, on average, to retain about 80% of the value they create. Overall, our results are informative about the workings of the managerial labor market.


You have to have mighty big faith in the strong EMH on studies like this.

dirk: maybe you could test your anti-EMH theory by going long on stocks that suffer sudden CEO deaths...

I haven't read the paper, but:

(1) I would think maybe there could be drops in value due to sudden changes in management, regardless of who was/is managing.

(2) That paper has no graphs at all!

From the paper: "We also observe that stock reactions on average become positive from day +2 to +3, which tend to be the days during which the firms nominate the interim executive or replacement."

Well that violates the strong EMH right there! The authors' have uncovered a predictive pattern in the market! So their own findings refute their premise.

I agree with @dirk. You are smart.

Why does this paper require belief in a strong version of the EMH?

Moreover, I don't think every version of the EMH implies that no prediction can ever be made. Predicting that a stock will rise by some unspecified amount is not much of a prediction anyway. After all, the S&P has an upward trend over time. It's like predicting that the economy will grow in the year 2011. It tends to grow at 3% a year or so.

Now, if the paper said something like--You will make a 10% riskless return in the 2 days after a CEO is found--well, that would be more dubious. But the claim that your cite is fairly unsurprising.

What does this say about executive life insurance and how to underwrite them.

What portfolio should the insurance company have to counteract or hedge a risk based on this finding. (By the way, any health studies of CEOs have to distinguish between CEOs who inherited the position from dad, and those who climbed to the top on a publicly traded organization. Stress levels are likely different for both groups, wealthy kids inheriting the business may have better healthcare than some schmuck who clawed to the top, etc.)

Most of the studies I've seen on stress show that those at the top of the pyramid have less stress and live longer because they can control their environment. These were monkey studies, so maybe we need to feed executives more bananas and less green.

Doesn't this argument overlook the strength of the successor? Suppose the CEO has put together an excellent management team and a succession plan. Then the CEO's sudden death should reduce firm value less than otherwise. Does that mean his value was less? I don't think so.

Of course the whole notion of CEO's being compensated according to their contribution to shareholder value merits a lot of examination. How many CEO's pay their companies back when the stock price drops?

Shouldn't this study take into account longer term price changes? The share price may well drop over their very short window due to uncertainty over the future. But suppose the negative returns are reversed in a month or two?

What would that tell us about the value of the CEO? Only that the new one is just as good. Fundamentally what happens is a brief panic, followed by a return to calm. The size of the panic may tell us what investors thought the value of the CEO was, but subsequent events reveal the truth.

Couple comments on paper, which i found interesting, but skimmed in 4 minutes.

1. odd paper only found poor returns on the day before, the day of, and the day after sudden deaths. They should look at total abnormal returns one month, one year, three years, and five years after the event. Also need to ask the question: Do markets over-react to sudden deaths?

2. 6 out of 150 of the observations were suicides. I wonder how many CEOs commit suicide right before a record earnings release... Heart attacks too, could well be endogenous.

3. compensation doesn't include perks -- catered lunches, corporate jets, company car (or does it?)

4. 67 out of 149 of the observations had the wrong sign. That's already perilously close to half, especially if we drop the suicides. Sounds like a coin flip with a smattering of large negative reactions also thrown in.

5. The interpretation of the stock market's reaction to CEO deaths as being indicative of a CEOs worth is problematic on several grounds. (a) that's just the market's perception, in actuality it's really, really difficult for anyone to objectively measure a CEOs value, except insiders, who cannot trade freely (b) A CEO death could also be a bad market signal, especially for the suicides, (c) as other commenters have noticed, a sudden, unexpected change in CEO should also get a negative market reaction, even if one assumes that the next CEO will be just as good, (d) If a CEO dies and the firm loses $18 million in value permanently, then that's not how much the CEO was worth, it's how much he was worth over what he was paid above and beyond the next CEO, who, if receiving the same salary, must be overpaid.

And, just found a red flag on Table 3: the cumulative return from the day prior to the day after death is -1.22%. That's a big drop. However, starting three days before to three days after changes things quite a big -- there's just a drop of just -.09%, which is well within the MoE of a null finding.

I think you guys need to update your thesis. It's problematic when you set out to prove some right-wing crazy-talk like "CEOs are paid their marginal product" when it's quite clearly far fetched. And null findings are important too.

I look forward to reading an updated draft!


Thorstein, Your comments sound persuasive, particularly re measurement of stock changes following deaths. I can imagine that stock changes could relate to the transactions costs of finding a new exec, paying the estate of the deceased, or uncooordination/fighting disruptions resulting from an unplanned change. But, I would be interested if there were the same results if the company had a succession plan in place v. one that didn't/.

Greg: My theory about how the firm works is that owners (shareholders) decide (indirectly through the board of directors for public companies) how much to pay the CEO. Thus, they are implicitly valuing the CEO's marginal "value creation" for the firm as being at least as high as his salary (or else paying him his salary reduces firm value and they ought not pay him that much).

I don't have much of a theory about how markets work, but to the extent that I do, if you believe in market efficiency then the CEO dieing reduces firm value because his positive marginal contribution to marginal value will no longer be made. A more cynical interpretation of the paper's results would be that shareholders wrongly believe CEOs to be relevant to the business. As a result, they react strongly to the death of the CEO, even though this is not a rational response. I will confess to not having read the paper, so maybe they discuss and respond to this criticism. It does not seem terribly novel to me, but maybe it is.



As I have said many times, the market isn't efficient, it is efficiency seeking (or creating).

Comments for this post are closed