« When should you buy a home? | Main | Avian Flu - update »
Facts about CEOs
1. In Danish data, if a CEO's child dies, the value of that CEO's company falls by one-fifth in the following two years.
2. If a CEO's wife dies, the value of that CEO's company falls by fifteen percent.
3. If a CEO's mother-in-law dies, the value of that CEO's company rises slightly.
4. American CEOs with McMansions run companies which significantly underperform the market
The Danish paper is here, the McMansions paper is here. On both studies, see today's WSJ, "Scholars Link Success of Firms to Lives of CEOs," the ungated link is here.
Posted by Tyler Cowen on September 5, 2007 at 08:19 AM in Economics | Permalink
Comments
I think your use of the word McMansions is imprecise. The WSJ article and the abstract for the paper made it sound like it was the purchase of huge "mega-mansions" that signaled underperformance. For a CEO, a McMansion would be a modest purchase in comparison.
Posted by: RWB at Sep 5, 2007 9:25:09 AM
followup to RWB: from the abrstract -
"future company performance is inversely related to the CEO's liquidation of company shares and options for financing the transaction... regardless of the source of finance, future company performance deteriorates when CEOs acquire extremely large or costly mansions and estates"
Posted by: jgray at Sep 5, 2007 9:41:15 AM
My favorite quote from the WSJ article (regarding privacy):
"I find it hard to imagine if I had a sick child that would be anybody's business," says Jerry W. Levin, chairman of Sharper Image Corp. and former CEO of Revlon Inc. "To assume that because something is going on in my personal life it's going to affect my business -- it's crazy. I wouldn't even ask those kinds of questions about my own employees, my own executives."
Certainly people are entitled to privacy. But, to suggest that the death of a child would *not* affect someone's performance... *that's* crazy. I know it certainly would affect my own job performance, and for that reason I might take a temporary leave of absence. What kind of person would you need to be for it not to affect your performance?
Posted by: Henry V at Sep 5, 2007 9:54:40 AM
henry...there's a difference there...he's saying that a SICK child wouldn't affect business. You're saying "how could a DEAD child NOT affect business?". Two different animals, there.
Posted by: shawn at Sep 5, 2007 10:12:57 AM
Nitpick: It's kind of annoying to have 1/5, 15%, and slightly used to describe the same variable, which I might want to compare mentally. Why not put them all in the same notation, like 20%, 15%, <5% (or whatever the right value is)?
I wonder how different the impact is for a sudden death vs. a slow death. It seems intuitively like those would be pretty different--if your wife dies suddenly in a car wreck vs. if she spends a couple years in a losing battle with cancer.
Posted by: albatross at Sep 5, 2007 10:25:48 AM
RWB is right. a McMansion is a cookie cutter big house on a small lot bought by the tasteless mass affluent with net worth no higher than the low millions, not the really rich. the study speaks of megamansions, something entirely different.
Posted by: Dylan at Sep 5, 2007 10:58:29 AM
Thanks for the correction on the wording. You can all infer, correctly, that we do not live in a McMansion.
Posted by: Tyler Cowen at Sep 5, 2007 11:18:29 AM
I don't know what a "McMansion" is, and this doesn't help:
"a McMansion is a cookie cutter big house on a small lot bought by the tasteless mass affluent with net worth no higher than the low millions, not the really rich."
The aesthetic value judgments there are so thick they can be cut.
So, it's a big house that you don't like?
Posted by: shawn at Sep 5, 2007 11:46:33 AM
3b. and if the CEO dies ... ?
Posted by: R at Sep 5, 2007 11:54:58 AM
A McMansion typically is built in a subdivision with other homes by the same builder and with very similar characteristics or styles. There is nothing wrong with them but their building styles tend to make them more affordable and easier to build.
Posted by: Steve Roberts at Sep 5, 2007 12:13:32 PM
This data should (but probably won't) show up in debates over CEO pay, specifically whether performance of firms is tied tightly enough to CEO performance that it can make outsized compensation packages a profitable move.
Posted by: jeh at Sep 5, 2007 12:16:32 PM
For the effect of a CEO's personal life on the firm to mean something more interesting it would be necessary to compare the effect in different countries.
Perhaps firms in Denmark tend to be more closely tied to the CEO, either because he is one of the founders or because of the size of the firm, or because of the way the management of the firm is structured.
In the US, CEO's tend to be fairly interchangeable these days and it is rare for their tenure to exceed five years. There are some notable exceptions such as chain-saw Al and Neutron Jack, but in general a change of CEO doesn't seem to do much over the long term. This is one of the criticisms of American CEO's they seem to be more interested in feathering their nest and getting out quickly rather than running the firm for the long-term benefit of all the stakeholders.
Another useful comparison would be with Japan where top decisions tend to be much more based upon consensus and not as dependent on the American Superstar model.
It's seem a bit odd that most political philosophers support the ideas of a democratic form of government, but none of them comment on the fact that all private firms are run as a type of dictatorship. Even the (nominal) control of a board of directors means nothing and the ability of the stockholders to replace them is essentially impossible.
Posted by: robertdfeinman at Sep 5, 2007 12:39:56 PM
Most of these effects are negative: the company underperforms when the the CEO is distracted. It is possible to argue that this shows that CEOs are extremely valuable to the company, but it could just as well be argued that it shows that CEOs have too much influence on their companies.
Everyone has probably some experience with a boss who wants to be involved in every project around, but is too busy to actually look into stuff. I can easily imagine such a scenario causing results like this: plans are made for important decisions, say a new factory or a reorganization, and the CEO says "let's wait a few months, I don't think the organization can handle it at this moment"
Although I have to say these numbers are very, very high. If 20% value loss is the average result, imagine a CEO who is above-average ingrained in the company(say he or she has been on the job for many years), and is above-average attached to the child(say it's the only one).
Posted by: Marius at Sep 5, 2007 12:54:32 PM
Is any of this enough to enable out performance of index funds?
Posted by: Floccina at Sep 5, 2007 2:09:39 PM
R: yes, from the abstract, CEO death also hurts companies.
The paper's take on this is very different from the WSJ article. The title is "Do CEO's matter", and the answer is that yes, they do, as measured by death or distraction of the CEO. So the idea is that a leave of absence won't help, because the replacement CEO won't be as good.
Posted by: DK at Sep 5, 2007 3:20:36 PM
"There is nothing wrong with [McMansions] but their building styles tend to make them more affordable and easier to build."
Uh. Maybe more affordable and easier to build than mansions of similar volume built individually with classic mansion-quality materials and on mansion-appropriate lots. But the typical starter castle is quite a bit more expensive to build, buy and own than a conventional house of equivalent (other than conspicuous consumption) function.
One relative who lives (in a conventional house) in a neighborhood chock full of McMansion-style homes reports, for example, that some of them cost $2K a month and up just for heating and air conditioning...
Posted by: paul at Sep 5, 2007 3:46:15 PM
I don't believe this for a second. The vast majority of a stock's performance is determined by its industry or asset category.
Posted by: bjk at Sep 5, 2007 4:38:19 PM
I'd always assumed the name "McMansions" was meant to echo "McDonalds," referring to the large, showy houses that are mass-produced for the relatively affluent masses from standard plans, like so many hamburgers.
Posted by: ed at Sep 5, 2007 4:41:57 PM
I don't think one can claim that proof that CEO's can negatively impact corporate performance is the same that they produce very positive results. It is much easier to screw things up than produce excellence. I still believe that most CEOs will produce around the same results. Obviously personal tragedy will impair that individual CEO and the company will suffer from the result. But absent that externality, most candidates for a CEO position will possess the same skill sets that allow for an expected return.
As for the Mega Mansions, it can be fun to speculate as to the reasons for the correlation. One could be that some CEO's see the company as their personal trough to pillage (through perks and compensation) as opposed to actually running the company. Mega Mansions, then, simply represent one of many forms of excess. Or it could be that at some point CEOs reach the level of compensation they desired and begin to slack off from work. Here Mega Mansions would be an indicator that the CEO has no further incentive to keep running a tight ship - he has no where to go, so he only wants to enjoy the ride now.
Posted by: Chris Durnell at Sep 5, 2007 5:41:00 PM
Ooops, I thought I was just being clever...
Posted by: R at Sep 5, 2007 6:01:52 PM
The certainty of some commenters that CEOs don't matter, or that they can only matter by making companies underperform, is fascinating. It is also absurd.
Aside from the empirical evidence presented here, there are strong theoretical reasons for believing that CEO quality is important. For instance, even if the CEO's only job was to pick his own direct reports, who in turn picked their subordinates, and so on, decisions made at the apex of the organization would have tremendous leverage all the way down the hierarchy. A good or bad personnel decision by the CEO would change the quality of hiring at the next level and so on down the line. That's not even looking at decisions about corproate boundaries, such acquisitons and divestitures, or other strategic choices where the CEO has a disproportionate influence.
Posted by: srp at Sep 5, 2007 6:41:38 PM
Is the McMansion effect very different from the jinxes and curses associated with celebrities being featured on magazine covers? These jinxes are said to be illusions, and are explained in terms of regression to the mean -- the celebrity appears on the magazine cover at a dramatic high point in his career, and afterwards, some downhill slide is only to be expected. When do these executives buy their mansions? I would think that in most instances, it would be right after a year of unusually successful business performance.
Posted by: Michael S. Gross at Sep 5, 2007 7:28:37 PM
AAPL holders: monitor Steve Jobs' personal life *very* carefully.
Posted by: fustercluck at Sep 5, 2007 10:37:09 PM
srp,
Don't most CEOs inherit the bulk of their directs (and their directs on down the line)? I'm sure that the hiring of any new CEO means that shuffling will occur, including some replacements high in the chain, but I think you're attributing undue influence on the typical CEO situation.
With a start-up situation though, I completely agree with you.
Posted by: fustercluck at Sep 5, 2007 10:41:24 PM
Or it could be that at some point CEOs reach the level of compensation they desired and begin to slack off from work.
IOW, they are overpaid.
Posted by: Disputo at Sep 5, 2007 11:56:35 PM
