Hacker News new | past | comments | ask | show | jobs | submit login
The MBA Myth and the Cult of the CEO (institutionalinvestor.com)
375 points by vinnyglennon on March 3, 2019 | hide | past | favorite | 159 comments

This is great work. It's a real job gathering and analysing all that CEO performance data.

My guess is that most people were suspecting this anyway. The fact is the CEO is there for a few years making a few key decisions, a large company is like an oil tanker, and markets are quite hard to predict.

Taleb has cooked a fair bit of soup on the fooled by randomness idea, but in this case his observations are correct. It is very very hard to show that CEO performance is something other than random.

That's not to say that everyone's performance is pure chance. Sports generates a lot of reproducible data about its performers, and we can tell from the numbers which players are better.

This is why it's not such a big deal for most people that Ronaldo or Messi get paid a lot more than their teammates, and the rest of society.

What does bother people, including myself, is that people are given far more credit for outcomes than is credible. Why should there be a many-multiples reward for being slightly up the senior management ladder at each step? Is it not most plausible that the leaders just under CEO also understand the firm's situation, and can make sensible decisions? And the step under them? Is it really so hard to find a CEO that you have to pay hundreds of grunt salaries to get one?

The other issue is MBAs. I don't think anyone is surprised that it doesn't help at all. I went to a well known business school too, and it's not like they tell you how to run a company. It's really just a kind of brainstorming exercise where you have bit of a think about what a company might do to be successful. What it does seem to be is a kind of stamp of ambition. People who go to business schools are keen to work hard to reach the top of the business world.

This might seem fine, but you have to wonder whether the best managers are ambitious people who want to be managers, or people in the line of business who want to compete for the same jobs.

I know who I'd pick.

What most long-term investors would prefer is a cult of a 'process', rather than a cult of 'personality'.

A process' is a combination of know-how + culture + rules + compliance.

Every business goes through cycles (eg startup, growth, pivot, then repeat)

At every cycle, different personal qualities of a CEO would be more beneficial.

However, a property of a company with a cult of 'process', is that the process will organically select and retain a right CEO at a right cycle.

So two outcomes of the above position: a) recognizing a company with a good process, is a very important and powerful skill to have (likely Warren Buffet has it).

b) averaging CEO performance without recognizing what cycle the company is at, compared to CEOs skills, is not by itself a measure of CEO performance. It is more of a proxy measure to whether the company has a sustainable and investment-ready process or not.


By the way, with regards to > This is why it's not such a big deal for most people that Ronaldo or Messi get paid a lot more than their teammates, and the rest of society.

I would challenge that, and say that it is quite abnormal, historically, for entertainers to get paid this much.

The development of multiplicative technologies (radio, TV, internet), and lack of wars between the 'super-nations' -- is really what made entertainers into exceptionally wealthy profession.

So, yes it reasonable to assume better performing sportman will get paid more than their team mates. But it is not reasonable that they get paid this much... that bothers me.

Athletes and entertainers being paid insane amounts isn’t as new as you might think, although fashions do wax and wane.


Gaius Appuleius Diocles was born in approximately 104 A.D in Lamecum... Diocles is believed to have started racing at the age of 18 in Ilerda... Records show that he won 1,462 out of the 4,257 four-horse races he competed in and was placed in an additional 1,438 races (mostly finishing in second place). The ‘champion of charioteers’ is one of the best-documented ancient athletes, most likely because he was such a star at the famous Roman Circus Maximus. [1] Being the best in the field also seems to have allowed Diocles to perfect his showmanship. Many of his victories took the form of a ‘come from behind’ crossing of the finish line at the last possible moment. The crowds loved it. Any race with Diocles quickly became the ‘featured event’ of the day. This naturally helped Diocles make even more money. His winnings reportedly totaled 35,863,120 sesterces, equivalent to 358,631.20 gold aureus or 2,600 kg of gold, allegedly, over $15 billion in today’s dollars, an amount which could provide a year's supply of grain to the entire city of Rome, or pay the Roman army at its height for a fifth of a year. Classics professor Peter Struck describes him as "the best paid athlete of all time".

Records show that he won 1,462 out of the 4,257 four-horse races he competed in and was placed in an additional 1,438 races (mostly finishing in second place).

>Many of his victories took the form of a ‘come from behind’ crossing of the finish line at the last possible moment. The crowds loved it.

What a wonderful story. The fact that he frequently finished in second place makes me think he was often trying to pull off a come from behind victory without quite managing it.

I can see how being instrumental in creating an exciting and tight race might be more lucrative than merely winning all the time. At least when the main source of revenue is spectators rather than sponsorship.

I'm not sure that a cult of process would be desirable - even putting aside irrationality of investors in the first place (and there is plenty in that field both mostly with short-term but also with long-term).

Process dominated companies often have issues with failing to adapt to a changing world as the reasons behind the process no longer exist and yet they stick with it anyway only to wind up out-competed and irrelevant.

On another note CEOs are generally overcompensated - especially if they wind up getting paid well even if they run things into the ground but I'm not sure process is a good antidote - I suspect the true issue is the lack of good feedback mechanisms/incentives given how many times companies have engaged in utterly stupid moves to try to inflate stock prices or earnings temporarily.

Meanwhile higher paid entertainers actually makes sense and are rational - they actually need to command the value fundamentally to be able to sustainably receive that much pay and paying them a 'reasonable' wage that massively undervalues them gives a strong incentive to jump ship when the enterprise depends upon them. It may not speak well to people's priorities given more important jobs that are paid far less but that is the economic reality.

Thx for the note.

So, I tend to separate a 'process' from a 'ritual'.

Most organizations would have 'rituals', but not processes in the context I noted above.

A 'ritual', is a process without clear goals, and without compliance and compensatory controls.

So, when a company has a pow-wow annual product strategy meeting, driven by a new CEO... or a technology group holds bi-monthly Architecture board meeting, or a data governance meeting -- those are mostly rituals. Not processes.

Rituals, masked as 'processes' are bad, of course.

They take oxygen out of the cult-of-the-process and move it back, again, to the cult-of-personality.

Which, then creates very similar 'profile' to the one of an organization that depends on a 'hero/mover and shaker/brilliant CEO'.

I do challenge myself at times, to compare how my disdain for the culture-of-personality would work at a war time..

Do general win wars, or do the troops?

I am going to say, at I need to learn quite a bit of more historical facts before, I can approach that subject.

Which is also why, I believe, solid bias-free education in history is almost essential for the work in management consulting, large organization transformations, and even creative consumer-oriented product design.


WRT my point on entertainers.

Their talents today, are not fundamentally different than the same talents that were appreciated 3000 years ago. Because these are bound to physical human abilities (and I include comedian/actors there as well).

This is in stark contrast to the professions that leverage scientific knowledge to achieve their expected outcomes (eg engineers, medical doctors, etc).

There is no net-positive-to-society economic model, where we can justify an actor making more than 1000 science teachers or a 1000 soldiers.

As I noted above, I think the entertainers benefited greatly from technologies, that allowed them to multiply and extend their performances in space and time.

For non-sportsman entertainers , additionally the copyright system , put them on the same footing as engineers leveraging the patent system (both systems are now heavily abused to install mafia-like control over engineering or entertainment content).

Lack of wars among the super-nations, and using entertainers as vehicles for political agenda, also gave them additional compensation leverage. Last 50 years, in western societies, have been exceptional for those professions.

And I find their current compensations whether it is Messi or Bill Maher, or Tom Cruise -- just repulsive (and even more so, than of mediocre or even talented CEOs).

> So, yes it reasonable to assume better performing sportman will get paid more than their team mates. But it is not reasonable that they get paid this much... that bothers me.

Would you prefer that the money just went to the owners, instead of the players on the field?

I think his view comes from a 'comparables' mindset. If you are to compare everyone else in the economy to entertainers, should they be receiving the multiples that they get? It makes other questions pop-up, for example.

If all the entertainer has to do is to practice and show up for the event, should he be making all of that or should a larger portion be distributed to all the other key variables that make him the center of attention?

> I would challenge that, and say that it is quite abnormal, historically, for entertainers to get paid this much.

Take a look at the salaries (certain) movie stars made in the 1930s

We don't need Presidents and Popes either.

The world is a much more complicated place today, than during the time of kings, and the accumulating data is going to keep revealing the complexity is too much for one man or woman to be responsible for it all.

But when a Deepwater Horizon type event happens and the whole herd starts baying for blood, are we sophisticated enough to say hey look at the data whats the point of lynching one person?

Of late the US seems to have attained that level of maturity, given how they have dealt up with Wall St after 2008, or the Pentagon for the 20 year war or dear Mark Zuckerberg. So maybe it's time.

I don't see older firms changing but newer firms might turn less hierarchical.

:Of late the US seems to have attained that level of maturity, given how they have dealt up with Wall St after 2008, or the Pentagon for the 20 year war or dear Mark Zuckerberg. So maybe it's time."

I don't think 2008 was a sign of maturity when the people who caused and benefitted from the meltdown got bailed out and could keep their money. Maturity would have been that the people who ran the banks before the crash would have lost all their money since it turned out that their contribution was overall negative. We have reached a point where the people who make the most money can hide behind big faceless corporations. They reap most of the profits while the losses go to the rest.

It's an interesting thought. Object oriented responsibility is as archiac as object oriented data analysis because since the rise of ML we know that the larger objects were just mental models for human decision making. However they erred by covering through averaging of more granular inputs that had conflicting information. I.e. if I'm a CEO and 3 of my VPs are good but one is terrible is it my fault that only one has become a risk. Under current object oriented thinking, yes but with granular analysis no it's the VPs fault.

Neither the President nor the Pope have a considerable amount of power.

The office of the pres. was specifically designed to 'not be a king' and he's highly constrained by the other wings of government. The US was supposed to mostly about the states.

The Catholic Church generally gives a very wide berth as to the operational artifacts of Parishes. He's also to some extent a figurehead.

CEO means such different things at different scale, at different organizations, it's really hard to compare.

I suggest the issues may have a lot to do with short-term vs. long-term incentives.

Also - 'average' pay isn't so helpful when there are a few with massive, massive pay packages that skew all of the data.

> Neither the President nor the Pope have a considerable amount of power.

Depends on which president you're talking about. The presidents of France and the US have a lot of power, the president of Germany has practically none. Monarchs in modern democracies often have less power than the German president.

And while presidents are highly constrained by other branches of the government, many pre-democracy kings didn't have any sort of absolute power either. Ultimately of course, any ruler can only rule with the consent of the people he rules; the real difference is to what extent those limitations are formalised in a law, like the US Constitution or the Magna Carta, and how far those limitations go.

In many companies, both the real source of a CEO's power and the real restrictions on his power come from outside the company, not from other employees of that company. The outside law supports the CEO's right to dictate how the company will be run, but also limits how much he can oppress or exploit his workers.

To play devil's advocate: The primary function of a manager is to say "No". As one goes higher and higher up the hierarchy, one has to say "No" to pushier and pushier subordinates, and the cost of every misplaced "Yes" is higher and higher. While leaders just under the CEO may understand the firm situation just fine, they may also be unable to say "No" to enough bad ideas for a myriad reasons, including lack of political clout with other leaders, with investors or with peer CEOs.

> The primary function of a manager is to say "No".

I'm a manager, and I can't really see this being the key job description.

But if it was, why would that have gotten harder in the last few decades?

Lots of people come to my team asking for work on X. If I said yes to everything my team would collapse from being spread too thinly. Being able to say that X is lower priority than what I am currently working on and leave the door open for re-evaluating priorities if needed is a key part of my job.

My team also frequently comes to me with ideas for interesting work that also isn't contributing to our vision. If we just did everything that came to mind we'd also fall apart.

My experience after 35 years is nothing scares me more than when a team of out of control developers are adding more and more features to the spec. Most baffling is it's obvious they think they are accomplishing something and you're a meanie for trying to rein them in. It always ends the same way. A lot of things can go wrong in a project. But that's that one 100% guaranteed to sink it.

On the flip side, a team of developers who aren't even interested in adding anything of their own volition is also a bad sign.

There has to be a balance.

Can’t say no to everything, or soon you won’t have anything to say no to any more.

You can't say "no" to everything, otherwise you'd be taking security's job.

In a good profession relationship 95% of the time proposals/issues are discussed until some agreement is reached or a decision to punt is made. The basis of that is a mature understanding that resources are limited.

Agreed. Being a decision maker means saying "yes" and "no", with the awareness that saying "yes" to some things, means saying "no" to others.

In my experience, those who say "the primacy function of manager is to say no" have either failed at management, or are working in organizations so conservative that they are immobile.

I think it varies with the group being managed. I've managed groups where many other groups hoped to benefit from our time, resources, and work products. In such groups a manager really does need to say "no" dozens of times for every "yes".

I think the primary function of a manager is to close the loop. Taking information from in and outside the group and turning that into a set of clear manageable requirements.

The job description is prioritization which is essentially saying "yes" or "no" to certain tasks. Especially in tech companies, the front line reports can always find something to do. Software developers want to automate everything or build cool things or use the latest tech. Managers are there to guide that, and that means saying "focus on A, B and C" because there's a gigantic list of tasks but saying no to 99% of that list makes the company perform better.

Next study "is ability to say no significantly different between COOs/CFOs and CEOs?"

My money is on no.

Saying “no” is easy. But it’s hard to say “yes” if it involves some risk.

See the movie industry

That's...a very apt example.

That is not primary function of manager. Misplaced no can be as disastrous as misplaced yes. Or just as irrelevant.

Also, the higher you go the more vague those people get, frequently saying neither nor yes.

Not to say “no” but to understand the company’s direction and only say “yes” to things that align. Important in this mix is communicating direction to the team and explaining the yes & no decisions.

The disastrous pivot that the Provident Financial did in the UK is one example.

Well, poor outcomes can certainly be attributed to poor direction from a CEO. Many of us have experienced that. My example is a new ceo who came in making a lot of noise about cleaning house and righting the ship. His intention was the sales org, but effectively scared off the best developers. He did also gut the sales org, and sales never recovered.

So objectively it’s not out of the question the reverse could be true, that a ceo could influence a positive outcome, maybe just harder to measure?

But this doesn't test what another CEO would have done in the same situation.

All we can do is make a guess about how the average CEO would have performed. There's no way to really test it.

"Why should there be a many-multiples reward for being slightly up the senior management ladder at each step?"

Really suggests some sort of a ponzi scheme...

> What it does seem to be is a kind of stamp of ambition. People who go to business schools are keen to work hard to reach the top of the business world.

Same as most degree credentialism. People don't hire Harvard graduates because of all the valuable stuff they learned at Harvard, they hire them because they were smart and driven enough to get in to Harvard.

Or they hire them because they went to Harvard as well.

> I don't think anyone is surprised that it doesn't help at all.

I know you're purposely being sensationalist, but there is pretty incredible value in the networking by doing an MBA...which IMHO is actually what makes good CEOs - having strong networks of people around you.

Right, clearly this makes strongly networked CEO, but not a "good" CEO in terms of subsequent period share returns, as the research demonstrates.

The CEO's decisions matter. For the most obvious example, see Apple. A series of CEOs drove Apple to near bankruptcy, then Jobs became CEO. With the same employees, plant, equipment, culture, etc., he turned it into the largest company in the world.

He did something similar to Pixar.

Microsoft is another example. Same company, same staff, same products, 3 CEOs, stark difference in results.

Your examples show that CEOs "can" matter, but not that they "do" matter.

I think anyone would agree that there are plenty of CEOs who happened to get their jobs just as pipelines come through, the economy rebalances in their favor, long-term HRD plans come to fruition etc. but they get all the credit anyway

The man in charge always gets the credit.

I have some AMZN stock. I wouldn't trade Bezos for someone cheaper. Would you?

To my knowledge Bezos gets paid only 80 grand a year - so no.

That's salary, not compensation.

It’s pretty much both.


Companies this year are reporting their CEO pay in relation to the compensation of a median worker for the first time as a result of a 2015 rule mandated by the Dodd-Frank act. Amazon Chief Executive Jeff Bezos recorded total compensation of $1,681,840 in 2017, almost all of which is money paid for personal security — Bezos drew a salary of $81,840 and received no additional stock. Bezos’s compensation is 59 times the median Amazon worker’s.

He doesn’t need to draw a fat salary or get more stock, he already owns more than $110,000,000,000 in stock.

Bezos is a hardcore megabrain who lives in the future

I think he is atypical, and so not a good representative example

Apple was eaten from the inside by Next, it's definitely not Jobs alone, it's Next using Apple capital and brand.

Jobs created Next, too.

> This is why it's not such a big deal for most people that Ronaldo or Messi get paid a lot more than their teammates, and the rest of society.

This is not why, and has nothing to do with it. Players get huge wages, because they are the best on their teams, and the sports generate a lot of money. If it wasn't going to the players, it would go to the owners.

One of the best things to come out of so-called “big data” is the slow destruction of the human ego.

I read an interesting joke few months back (not sure where, might have been in Capital Minded), the gist of it was basically:

If 99% of people who buy corporations for a living can't beat a monkey throwing darts over time...why do we believe people who run corporations for a living are any smarter? They're both looking at the same corporate balance sheets and industry trends!

I mean, it makes intuitive sense. For a fully scaled company in a legacy industry, what would really change if we removed the CEO and replaced him with the Janitor? The business will still be subject to the greater trends of the market and its industry. It doesn't take a Harvard MBA to take a data input like: electric cars are the fastest growing automotive segment therefore...my auto company should probably invest in electric cars?

Edit: where I think this could fall apart is in frontier industries like tech. The decisions require much more specialized expertise and the industries are much less mature (ie. Say ridesharing).

“If 99% of people who buy corporations for a living can't beat a monkey throwing darts over time...why do we believe people who run corporations for a living are any smarter?“

Because the 1st requires predicting what other investors think about a company or the economy (after all, the stock price doesn’t move in align with its profits/revenues). Not whether a company will be successful.

The 2nd simply requires knowing what consumers want, whether they be individuals or businesses. You just need to solve their problems in a better way than competitors. So understanding the fundamentals of a market is enough.

What you're describing is the "Greater Fool Theory" of asset pricing...where the emotional temperature of other investors sets the price for an asset. This is true only for assets that don't produce cash flows and thus can't be valued using traditional valuation models (ie. like gold, crypto, etc).

Businesses on the other hand are properly valued for their profit/revenue potential given the available information at the time. This has been proven over and over again. The data is clear on this; the high level of market efficiency is why the stock market is so hard to beat.

Successful investing (not trading) absolutely does require predicting whether a company will be successful. Both investors and employees of a company need to be correct about the thesis for why the company solves a consumer need. Otherwise both parties will lose.

IIRC, this is one of Warren Buffets main strategies and why he beats the market. He mainly invests in companies that are undervalued and with good cash flows. He doesn't have to risk the market coming to their senses like with shorting because even if the stock price goes down, he just holds onto it and rakes in those cash flows.

Point taken. I was talking more about trading, not investing long-term. If we're talking about long-term investing, most funds have strict limitations, and can't be too focused on any one single stock/sector too much.

Warren Buffet famously says he likes to invest in businesses that could be run by a ham sandwich.

"I try to invest in businesses that are so wonderful that an idiot can run them.

Because sooner or later, one will." --Warren Buffett

Whether the quote is by Buffet or not, I really wish I could be so "idiotic" as to get paid tens of millions of dollars a year as a CEO.

Just one year would be enough, actually.

I think the quote is by Peter Lynch from the book - One Up on Wall Street.

Perhaps, though Googling the quote has SERPs more than 25:1 in favor of Buffett as origin.

No wonder he's doubling down on Apple.

> what would really change if we removed the CEO and replaced him with the Janitor?

I think this reads rather more out of the conclusion than the data can bear. It's possible (though I don't think plausible) that it's true, but the data is only based on people who were selected to run companies i.e. people who those doing the hiring saw to be a good fit. It says nothing about being run by a random person off the street, for which there is necessarily no data.

> If 99% of people who buy corporations for a living can't beat a monkey throwing darts over time...why do we believe people who run corporations for a living are any smarter? They're both looking at the same corporate balance sheets and industry trends!

The market acts as a filter which means that people still need to create and run create the enterprises for it to be able to value them.

Anecdotally, having spoken with both CEOs of very big name billion+ dollar companies and with many money managers, professional investors, and various other finance hucksters, I can say with assurance that the high end CEOs have considerably greater cognitive ability on average. People like Warren Buffett represent a kind of best of both though, since he is a successful CEO as well as investor. I think it comes down competition. Managing someone else's money for a 1-2% vigorish isn't really an activity that takes much investing skill to be successful at, especially since nobody else knows how to predict dynamic chaotic systems either.

Also, given how many new businesses fail, there is probably some survivor bias here too on CEO performance.

An important part of an executive’s job is hiring the right people to execute, which requires some level of skill or networking in my opinion, even if results aren’t always guaranteed.

But the data here says there is no statistically significant benefit to picking an “in-network” (ie. Former McKinsey, Goldman Sachs, Harvard, etc) executive. Which is the homogenous network that these people will continue to cultivate.

This data says the parochial boys club doesn’t deliver any difference from the average state-school “rose through the ranks” type of executive.

The person I was responding to was postulating what would happen if a janitor was made CEO, to which I am saying there are some important skills or experiences that not everyone will have.

I think they were being a bit extremist with their language to make a point. However, fun anecdote, at my high school one of the janitors was a multi-millionaire (we lived in a fly-over state in a small city, so he was crazy rich by our standards).

He made it all because he was extremely good at playing/investing in the stock market (he never had a professional job in his life). The economics teachers would invite him to speak to their classes on a regular basis.

In other words: access to finance requires membership in an old boys club.

> One of the best things to come out of so-called “big data” is the slow destruction of the human ego.

True! But this isn't a case of big data. This is just data. They only looked at 8500 CEOs and only did basic statistical analyses, ie not "big data".

One of the key balances we have to have with big data is to not let our personal biases and hubris blind us to believe in incomplete formulas/solutions simply because it follows our preconceived notion of ourselves, the tool (big data in this case), etc.

What if you look at non-frontier industries where utilizing frontier know-hows can create a competitive advantage?

> If 99% of people who buy corporations for a living can't beat a monkey throwing darts over time...

They're not "buying corporations", they're gambling. As such, their success rate is about the same as in other forms of gambling.

I've done these same analyses before as well, and more. This is the correct analysis to do if you're looking to build an alpha-generative trading strategy. It is not necessarily the correct analysis to do if you want to measure actual CEO performance.

If you want to measure CEO performance, you need to look at growth in fundamentals, not share price. The share price is going to reflect the market's beliefs about the CEOs. Which, if the market is efficient (as it seems to be for these variables), should show exactly no effect. If there were an effect here manifest in share price returns, we could all read this article and arbitrage it away. This would mean that the next person who looked for this effect would conclude "CEO performance has no correlation to share price returns!".

I'm kind of surprised they would do all of this statistical work and not be aware of this gaping hole in their reasoning. Now they do make the point at the end that they're arguing against stock-based compensation here, which is a good point. If the market prices in your value as a CEO the moment you're hired, you don't want to have share-price based compensation. You want comp tied to growth in fundamentals.

Did you read the article? The objective was to prove or disprove the Jensen thesis that CEO compensation should be tied to gains in stock price because CEOs affect stock price. They were NOT attempting to answer the question of whether or not CEO quality and compensation affects the success of their underlying businesses.

Did you? They were attempting to do both. I specifically called out that it was a legitimate way to do analyze the compensation issue.

That depends on how the effect is being measured - if the price growth is measured from the point where the CEO is appointed publicly - then you are right. But if they start from the moment before the announcement - then there would be no way to arbitrage the effect away. The interesting thing to do would be to measure the effect of the announcement itself.

Ya, I agree. However, i'm pretty sure they're doing it from the moment of appointment, not the moment of announcement. They don't call it out specifically in the article, but i'm familiar with the datasets they're using, and I don't believe announcement time is in there. Though it's certainly possible that they have data I haven't seen.

Thank you.

I also wish they would have measured startup (ie, private) CEO performance. Of course that's quite hard because financials are private and all the failed companies have no records left, and if they did, good luck teasing the data out.

> You want comp tied to growth in fundamentals.

This has the same problem as tying it to the market. You don't know what the baseline is. Would a good CEO always grow the fundamentals? What if it's a declining industry? And vice versa.

It is a problem, but it's not the same problem. You can try to address the issue of secular trends in the industry by looking at say, performance relative to industry peers. How you define industry peers of course is a bit tricky.

Another option is to look at share price returns from the day before the new CEO was announced. That would incorporate the market pricing in the new CEO. However, I don't think this analysis did that.

It's not the same problem because expectations of CEO outperformance can be internalised into the price at time 0.

Except, we've learned via a robust body of research over the last 5 decades that the market is largely efficient.

Therefore, a change in share price over time (not measured in quarters, but over 3+ year periods...as was done in this study) is an excellent proxy for change in fundamentals.

> Except, we've learned via a robust body of research over the last 5 decades that the market is largely efficient.

Oh? Care to cite some of that research? And let me preempt a few common citations which don't actually work:

1. Warren Buffett's hedge fund bet does not prove this. Among other reasons, he made a bet concerning the performance of a fund of funds, not any particular outlier.

2. Eugene Fama's research on the Efficient Market Hypothesis (EMH) does not prove this (while we're at it, Fama would disagree with you that the market is efficient). His work applies the EMH as a means of studying efficiency, not asserting efficiency. One of the most common misconceptions of Fama's work is that it's descriptive instead of prescriptive.

3. When you actually do the math, the number of hedge funds that has ever existed is nowhere near what you'd need to explain the outliers like RenTech or Baupost. So the common coin flipping analogy actually doesn't prove it either.

I can save you a trip to look for citations - the "robust" body of research you're talking about doesn't exist. There is no such consensus. The efficient market is a model which only approximates real world conditions - it do so imperfectly, and in the best case it only works locally.

Here is an article [1] written by Cliff Asness on the topic of the EMH which might be educational. Asness earned his PhD under Fama and founded AQR, one of the more successful hedge funds. He also regularly comments on, and published, research. There remains a great feal of inefficiency at the macro level, even on horizons measured in years.


1. https://www.institutionalinvestor.com/article/b14zbgrj5pflsc...

As I stated, the market is largely efficient. I did not say it is completely efficient.

Also throwing out Rentech as an example just further proves my point.

The inefficiencies that Medallion fund is exploiting are extremely capital constrained (ie. They are very tiny). This is why they don’t accept outside investor money and even limit the amounts employees can have in the fund. The fund is limited to roughly $3 billion, because the inefficiency they are exploiting is that small.

An inefficiency of $3 billion in capital markets of roughly $100 trillion dollars is tiny.

All of Rentech’s larger funds for outside investors have not shown the same ability to beat the market.

Also, I’m amazed that you use Asness as a counterpoint, because if you’ve ever heard him speak, he would agree that the market is “largely efficient.” I’ve literally heard him say those words in person.

Asness’s entire hedge fund is based on Fama’s factor models. He literally makes his money by making the market even more efficient. He has admitted long-short factor funds like his have likely reduced (and will continue to reduce) factor alpha, hence why they need to apply leverage to make these returns significant.

The market is going to price in expected change in fundamentals due to the new CEO. They're going to price it in immediately - before the fundamentals actually change. The market does not proxy change in fundamentals. It proxies change in expected future fundamentals. These are different things.

We could be arguing semantics here, but my point is; the market will of course price in prevailing sentiment at the time, but this will be corrected over the next 3 year of earnings releases whether prospects for future fundamentals have actually improved (hence why the study used 3-6 year time periods).

The price at year 1 will be based on the prospect of future earnings. The price at year 3 will still be based on the prospect of future earnings...but with 12 quarters of hard data allowing you to more accurately price in the new CEOs specific strategy and its effects on growth prospects.

Right. But any of that future price movement that was predictable based on the known attributes of the CEO at the time will be priced into the initial move. If it were not, then you would have a profitable trading strategy that say, went long Harvard CEOs and short Princeton CEOs (strictly a hypothetical example).

The lead author is a friend of mine. The beautiful thing is that he got an...MBA from Stanford.

I’m consistently intrigued by his observations about the world. He doesn’t let his own background get in the way of questioning established wisdom. The fund he runs now is based on a belief that private equity is seriously flawed...which he discovered by working as an analyst for one of the top firms.

He hosts a fantastic weekly newsletter with a lot of quantitative insights - can discover his strategy and sign up here:


And the management team is...all MBAs :)

>This email list is only open to accredited investors. By clicking "subscribe to this list," you are representing that you are an accredited investor.

Is that for his own protection? People aren't going to drop from helicopters into my home, are they?

I'm not sure this proves anything at all.

To play devil's advocate: what if top CEO's (whether by school or by track record) look for harder challenges?

Someone once said Dara Khosrowshahi took the CEO job of Uber not to make money, but because anyone willing to take that job is somehow who wants to play the CEO game on "hard mode".

If everyone's taking a job where there's a 50% chance of failure for their personal skill level -- which means the more skilled CEO's are taking the harder jobs, and the less skilled CEO's are taking the "safer" harder-to-mess-up jobs -- then statistics will look like CEO skills don't make a difference... even though they do.

And since employment markets (in general) tend to funnel higher performers to harder jobs (because that's economically efficient), this doesn't seem implausible.

>Over the past year, we set out to answer these questions. We created a database of approximately 8,500 CEOs and their characteristics, each individually mapped to their respective companies for the duration of their tenure, and pulled company fundamentals from Compustat, stock returns from the University of Chicago's Center for Research in Security Prices (CRSP), CEO tenure and education from BoardEx, and long-form CEO biographies from Capital IQ. We then ran a battery of tests on the new data set, looking for correlation, persistence, and predictive power.

...Nevermind that, I heard that some guy took the job because he liked challenges.

Maybe Dara was already wealthy and the marginal utility of the money at Uber wasn't a factor? Maybe this would lead him to beleive that his own decision making an knowledge would be superior and he wouldnt listen to his peers. Soooo many variables.

It never ceases to amaze me we can dismiss data driven ideas in a blink of an eye because it doesn't fit a narrative we've developed through anecdotes.

> It never ceases to amaze me we can dismiss data driven ideas in a blink of an eye because it doesn't fit a narrative we've developed through anecdotes.

This is perhaps because most people develop a paradigm of certain things and they're usually unwilling to accept a different one, even if the data is shouting at them to do so. And the more fundamental reason may be that people just don't like change.

You’re absolutely correct, some of the time.

On the other hand, at other times, academic researchers can conjure up an armada of buzz words and hot air balloons to support a position that even very limited real-world experience and common sense proves laughable misguided.

"common sense" and "real-world experience" are frequently euphemisms for "everyone knows that's how it works".

you just did exactly what I was talking about.

Yes it was very obvious that spontaneous generatuon was the correct answer, that Pasteur guy was a moron.

I have been proximal to the CEO search process at a few different large companies, and at a minimum have anecdotal insight into what drives selection.

CEOs are hired to fix perceived fundamental problems with the business, even at externally successful companies, and are evaluated for fit based on their history solving similar problems at other companies. Many times the business problem driving selection is invisible to casual observers, particularly those external to the company because it involves internal board/investor topics. The CEO's primary job when hired is fixing whatever it is the board deems it necessary to fix, not posting good numbers per se.

Fixing a company is exponentially more difficult the larger the company. The ability of most people to effectively reshape a company in some key way scales very poorly with company size. There is definitely a Peter Principle at work and most successful CEOs fail to be effective at some scale. To make matters worse, most boards have ambitions to grow the scale of the business, so they tend to want to hire CEOs with experience that exceeds the current scale of the business, which creates an intrinsic shortage of candidates with proven experience to reshape organizations at the desired scale. Supply and demand.

It is ironic that this site often has horror stories of working at startups -- tiny organizations -- with disastrous CEOs, but in the same breath people assert that anyone could be an effective CEO of a 10,000-person organization because the job entails no meaningful skill. Most people are demonstrably ineffective as CEOs in a 25-person organization, there is no reason to believe they would suddenly become effective if the organization was a hundred times as large.

> but in the same breath people assert that anyone could be an effective CEO of a 10,000-person organization because the job entails no meaningful skill

That is not the assertion being made. The assertion is that most CEO's cannot make demonstrable contributions to the growth of a fairly mature company.

Sounds like an unnecessarily complicated explanation for the observation the authors made.

There are two competing hypotheses: first, CEOs' CVs and track records impact their future success, but a market mechanism ensures they get promoted into jobs in just such a way that their impact is offset by the difficulty of the job. Second, CEOs skills are not measurable and their prior success does not predict their future success.

Occam's razor sort of suggests that the latter hypothesis is preferable.

I'm reminded of a recent 'paper' where the authors took the idea of meritocratic advancement out for a serious shellacking. Their point the winnowing effect doesn't move the needle vs competence that much.

Yep. Also this is all based on stock price data, so it could be that businesses typically hire the person that the market "expects"...which would predict zero alpha even if counterfactually hiring a bad CEO would've tanked the stock price.

As already mentioned by another commenter, freakonomics had a good series on CEOs, including an episode on the glass cliff: http://freakonomics.com/podcast/glass-cliff/

The hypothesis being: female/minority CEOs are more likely to be hired as CEOs at failing firms (perhaps) because male candidates are less likely to apply/accept. Which, if true, would contradict your hypothesis.

Couldn't we just broaden the definition - a good CEO takes a job that is difficult but achievable, a bad CEO takes a job that he's not competent enough to identify as impossible. I think a good example of this is Intel. Intel just spent a year trying to find a new CEO, but failed because no one qualified to do the job was incompetent enough to accept the job. All the candidates were competent enough to realise that job is going to be a blood bath.

> To play devil's advocate: what if top CEO's (whether by school or by track record) look for harder challenges?

That literally means that CEO's do not know how to do their jobs. Even if they do it knowing it.

> Someone once said Dara Khosrowshahi took the CEO job of Uber not to make money, but because anyone willing to take that job is somehow who wants to play the CEO game on "hard mode".

And there is families that depend on that companies salary. Drivers around the world will be affected by Uber's CEO decisions. That is a job that will not be a game for most people. But, maybe it is for anyone that want that kind of job. They are just playing while others are the ones that are affected by it.

I definitely think you are right. But then again, there are so many confounding factors at play that this all becomes so difficult to detangle.

I posted this in a separate comment too: Freakonomics Radio had a series called "The Secret Life of a CEO" that covered CEO performance and many other topics. I highly recommend it!


Not really. I would put it this way. What would you do when you have build a reputation over years. Would you simply forego it for an adventure. I don't think so and I have not seen example to prove your hypotheses.

Even when these guys have taken impossible jobs, they negotiated for themselves first and ensured getting paid even when they didn't succeed.

That does also suggest that rewarding with stock is not reasonable: if CEOs were actually motivated by money, they would go for easy mode instead.

He took the job of Uber CEO because he'll make 100x more money than doing anything else.

Taking over Uber was not 'hard mode' it's 'easy mode'.

'Hard' is building a business that scales and that's making money. That's extremely hard.

Uber has utterly fantastic fundamentals - they are growing rapidly and have a very powerful brand (despite their CEO shenanigans, riders don't really care, they'll take Uber if it's convenient).

As the previous CEO self destructed with some bad bits of PR, the 'new CEO's job is to come in and be the new face of the company as he prepares for the IPO.

The new CEO can blame the old CEO for all of the cultural problems, which are frankly not that hard to fix - after all it's mostly cosmetic. Fire a few really bad apples or those who can't be tolerated, launch a new social program, make it public etc.. The vast majority of Uber staff are just regular people.

Some restructuring would be required, such as axing some of the more aggressive secondary plays, which allows the nice unit-fundamentals to shine through.

And then do one of the biggest IPO's in history.

Taking over for a CEO who mostly has a 'perception problem' of a company that's growing like gangbusters and about to go IPO is most CEO's dream come true.

Even if "firm performance is not predicted by the educational background of the CEOs", it does not follow that an MBA is worthless.

To draw a parallel, I work with many software engineers both with and without CS degrees, and there seems to be no correlation (or even a slight negative correlation) between performance and those with a degree. That does not mean CS degrees are worthless! It means once you have already selected for folks with a successful career in software, the degree has less predictive capability.

Bingo. If you look at successful people you’re probably going to find a pretty diverse set of backgrounds in addition to the developers with CS majors from top schools. But doesn’t mean that spending your 20s as a ski bum is actually a recommended path for most professional careers—even if it does make for a cool story if you get to the point people are writing magazine articles about you.

I always figured to be an effective leader requires knowing your product. For software companies that means working in their field so you understand the tech or problem area then can make better decisions. You can complement your leadership by getting an MBA which rounds out the rough edges of business.

Career MBAs who have no experience of the area of the business are red flags in my experience. It's like a manager who isn't technical. They make wrong decisions by themselves or listen to the wrong people.

>> I always figured to be an effective leader requires knowing your product.

One of the best bosses I ever had was when I worked at a small manufacturing company.

This guy spent a ton of time with our customers and our products, came up with improvements, made lots of fixes, found other products our customers could use, made some people really happy.

Then he went through the production floor. Looked at everything a fresh set of eyes, why do we do this, why don't we do this. Sped up production, saved money, simplified things, made everyone's life easier.

Next stop was the office. Why do we touch this document 5 times through 4 people? Why do we do this backwards? How come this person doesn't do anything all day? Let's move this schedule around.

It was about that time the owners fired him for petty shit even though we had record sales and profits for the 3 years he was here. Lesson learned, don't try to make any changes or recommendations that come a little too close to home.

Interesting story. I'm curious about the "petty shit" part. Did he step on the owners' toes? Did he have an abrasive personality? Was he fudging his expense reports?

Also curious what happened after he left. Did sales/profits drop? Did his legacy of reforms carry on?

Things are starting to go back down, but they’re staying in the black at least for now.

Petty things would be exposing inefficiencies in the owner’s pet departments, usually consisting of underperforming employees that are family members, and the owners undermining random decisions he made just to spite him.

Not OP, but...

1. Manager insisted on proper documentation of contract deliverables in case of a lawsuit. Higher manager said “Nah, waste of time. These companies never sue.” Lawsuit happened, case was overcomplicated due to poor documentation. Result of case still unknown, but should have been a slam dunk win for the company.

2. CFO decided to delay (not reduce) part of a regular disbursement to the owner in order to pay suppliers. If suppliers had not been paid, then product line gets shut down and company has an existential crisis. Disbursement eventually got paid, suppliers were happy, company thrived. CFO permanently in the dog house (e.g., lower bonuses, lower pay increases, left out of key meetings, etc.).

3. Any whistleblower in the US (maybe not “petty”). While whistleblower protections exist, it’s trivially easy in white collar work to find minor mistakes or question judgment calls. Retaliation can be trivially masked. Usually business as usual afterwards.

4. Taking alloted and approved vacation time — this happens a lot. Usually ends up causing the company to underperform after star leaves. I’ve seen it wreck a company when the person who left recruited all of their good people to his new company.

5. Documented firing of disruptive sociopaths who are friends with a senior manager (e.g., the CEO). I’ve seen this multiple times. Overall org usually becomes better due to lack of sociopaths, but the “hero” gets the shaft in the process.

6. Tactfully bringing up issues of fraternization by senior managers that may lead to conflicts of interest. I’ve seen this several times as well. Simple case is a senior executive who is married is sleeping with a low level employee. More complex is someone who is in an open marriage who is using junior employees not in his current department (but may be in his future department since they often make lateral moves) as a personal dating/sex pool. It usually ends up as a non-issue, until it becomes an issue (e.g., Ellen Pao).

7. I’ve literally seen someone fired for streamlining their department so much that it appeared that they didn’t do much work. The person was aware of the fact, and pitched it as “maintaining the system”. The management structure was “flattened” after that person was let go, things went to shit, and no one in senior leadership seemed to know exactly why.

8. Worker (not manager) reviewing vendors said that none of the vendors who submitted bids were able to do the work to standard based on work samples. Worker was replaced, contract went out, company lost millions due to... contractor not being able to produce product to standard.

There are many more...

What edges does the MBA rough out?

I've always been skeptical of them, and my friends with MBAs haven't helped. They were all optimistic about how much they were going to learn, but by the end thought "at least I got a network and a degree on my resume".

I don't think anyone would get an MBA if you couldn't see it on their resume.

Whichever edges are rough for you personally, that's the point.

Like if you're great at finances but terrible at understanding sales and marketing, then the sales and marketing part of the MBA will be most valuable for you. And vice-versa, and all the other combinations.

MBA's ensure you have at least a minimal competence in each classic area of running a business, and aren't going to make bone-headed decisions that might otherwise seem logically plausible.

That's one of the reasons MBA programs tend to be centered around case-studies: so you're exposed to a ton of ideas that seemed reasonable but then turned out to be disastrous. As long as you've done the MBA, you're unlikely to repeat those. If you don't have an MBA, you might not know any better.

A Master of Business Administration does for the business what the System Administrator does for IT systems.

MBA is a fine discipline. The problem is when MBA gets equated with SVP/CEO.

> You can complement your leadership by getting an MBA which rounds out the rough edges of business.

If you're already a leader, an MBA can help.

However, I see a lot of MBAs who come skipping out of a B school (even a good one like Wharton or Kellogg), and their approaches to problem solving are so identical!

I have seen my share of CEOs, but it's staggering how similar they are in their approaches. You could, after interviewing a few of them, come up with a flow chart and you'd capture about 90% of their actions. I have seen many a CEO run the company into the ground, but very, very rarely seen a CEO lift a company up and make a difference.

I think Satya Nadella in MSFT has really turned that ship around (please correct me if I'm wrong). But for every Satya, there are a 1000 mediocre ones, just trying one tool after another that was taught them in B school, and when the board gets tired, it throws them out with a golden parachute, only to be replaced by another cookie-cutter who does the same.

But is CEO quality really about the best decisions?

Playing the MBA's advocate, maybe it is so hard to "consistently beat the market" with decisions, that their usual quality spread does not even matter that much? Amongst other reasons, because even an apparently foolish decision, if for some unexpected turn of events it suddenly is the "golden path" will give the bigger returns than an "obviously smart" decision ever could?

If that is the case then the decisions themselves would be secondary to how those decisions are executed. Then you wouldn't look for the best decisionmaker for filling the CEO position, you'd search for the best at getting the team united behind a decision who isn't completely terrible at making them. Basically someone who is good at commanding respect, which is exactly what those high-status institutions are teaching/selling.

Being the typical overconfident dev I always tend think that I would make the smartest decisions ever if in the captain's chair (surely this isn't true, but it feels true way too often), but I never think that this would make me an actually good CEO.

The basic unit of information for a human being is a story, not a number or a fact. So it doesn't surprise me that we're obsessed with something as trivial as how Elon Musk's $1/day diet shaped him into an entrepreneur[1].

I don't think it's a bad thing to have a vague awareness of the things that make people successful: unafraid of failure, hard work & dedication, true passion about what you're creating.

But I think it does actively hurt you if you overfit what causes success, because you focus way to much on getting the little things right that you miss the big picture.

[1] https://www.independent.co.uk/life-style/gadgets-and-tech/ne...

> unafraid of failure, hard work & dedication, true passion about what you're creating

I would say those are in equal measure things that make people unsuccessful. You better be afraid to fail or you will follow too many stupid ideas. You can burn yourself out working too hard on the wrong thing. You can be really passionate about something that basically nobody except you wants or needs. Conversely, a lot of things that nobody really is passionate about can make a lot of money.

I will also say that Elon Musk has not been successful in real business terms (creating a sustainable, profitable company) except that one time with PayPal and that could be down to good luck and timing. Nowadays, there's too much cheap and dumb money chasing too many "next big things", so it's hard to tell fact from fiction.

In my estimation, success really comes down to:

1) showing up

2) timing

3) luck

There's lots of people that are almost Elon Musk or almost Mark Zuckerberg or almost Steve Jobs. It's a gamble, but at least you get to choose which table to play on.

I don't find changes to share price to be the most convincing indicators of performance, though I agree you'd probably be hard pressed to find a better "generalizable" statistic. Fundamentally, I think measuring talent is an astonishingly hard problem.

My biggest takeaway from this is to invest in a company right when the CEO changes if they CEO came from a failed company. Assuming "strong track record" is priced into the stock of a company picking up a new CEO that's good, I also assume "weak track record" is priced in to the opposite. Since success afterwards is basically chance, it would be a discount.

Freakonomics Radio had a nice podcast series, The Secret Life of a CEO, that looked into the effect of CEOs on companies among other interesting stuff like how is a CEO chosen, what happens when they are fired, etc.


This is kind of crazy. Presumably, there is a correlation between going to a good business school and getting a job with a good business (which generates market-beating returns). So you would expect some kind of correlation just because people think MBAs are valuable...but no.

In particular, "Elite MBAs did perform relatively well as CEOs in healthcare and consumer staples". Not significant statistically but the reason this is true is because both healthcare and staples (until recently) have had a killer run.

So I don't think share price performance is the best metric but...even then...you would still expect to see something totally different to the actual results. The stuff that looks at the portfolios of bankers is more understandable...but this has been shown elsewhere (generally, poor fund managers do better).

The success of a company rests on thousands of factors one of which is the CEO. Even the best CEO can’t ensure success when all other factors are against it, but a mediocre CEO will sometimes fail when all favors are in their favor.

Which is to say. When trying to figure out if a CEO is ‘worth it’ don’t like at company success as a metric. You need to look across the industry and find “big obviusly avoidable but extremely costly mistakes” and see how much they cost, and ask: is it worth it if they can avoid all thoses issues.

I see a few critical problems with this research.

The first in the section "Do MBAs Make Better CEOs?". If the labour market is efficient we would NOT expect CEOs with MBAs to outperform CEOs without MBAs due to selection effects. We may expect them to outperform if we sample randomly from the general population that has MBAs versus that which does not, but this is not the sampling scheme in effect. If boards are able to accurately select non-MBA CEOs whose other characteristics compensate for the lack of an MBA, then we would expect a zero within-group correlation in the selected population. It is like if I was to hire a quantitative researcher for a machine learning research team. I could pick a PhD from a top school, or I could allow a non-PhD into my team as long as their other characteristics are sufficiently great (e.g. competition math in high school), but after appropriate selection effects there is no within-group correlation between holding a PhD and not holding one. If I looked at the average PhD versus the average non-PhD in the general population though, then an effect becomes apparent, but that's not the sampling scheme I've used.

The second problem is in the section "Is CEO Performance Persistent?". They used stock market returns which is flawed as expected CEO performance should be baked into the stock price from the beginning, meaning we expect zero excess return in the second sample even if the CEO was truly exceptional. This section would unlikely pass peer review into a top finance journal.

The only valid thing I can see is research about "Share Price Performance for CEOs Who Ran Multiple Companies", since nobody knows the second company ahead of time so the information shouldn't be baked into the price. But a lot of details are missing. Skewness and kurtosis of returns will impact the way they've quantized the data, among other things.

All in all, unpersuasive.

The design is interesting but the test itself is kinda pointless. Fama-French explains about 30% of stock return variation at best so by using stock performance as a measurement you are mostly testing signal (highly credited CEOs) against noise (residual returns).

Another thing is you have to specify your position on EMH here. If you believe the semi-strong form then everything about the CEO should be priced-in immediately after the announcement, which means it will not have a persistent impact on stock return when you slice and dice time periods.

IMO the CEO is just one input out of thousands that influence a stock's performance, which is implied by a lot of asset pricing studies already.

Also managing a public company is an unbelievably complicated job. You have to be one of the "best and the brightest" just to keep the boat bearly afloat sometimes.

When a management with a reputation for brilliance tackles a business with a reputation for bad economics, it is the reputation of the business that remains intact. - Warren Buffett

The business that the company is in matters much more than any individual CEO or their background.

I am thinking about starting MBA in the fall (having MS in CS from a Top 10 US school), just because I can and it could be helpful for running my own companies. Talk me out of it please!

(I might do a PhD in ML in parallel if I feel crazy enough)

It sounds like you should do... neither.

Your expectations seem to be very high and different from what the degree will deliver, on an average (you might be an outlier and still want to do it).

MBA: * expectations: help start your business * reality: networking, understand business process, opportunity cost (both time and money)

PhD: * expectations: help start your business * reality: petty politics, super high opportunity cost (both time and money), possibly minor improvement in research field (if that)

How would the MBA help in "starting & running a company"? I think the PhD may provide a better launchpad to starting something up. You may want to correlate startup founders to MBA degrees at the time of founding if you need hard data.

That's why I am asking because I have some doubts. But there are specific entrepreneur/innovation oriented MBAs oriented in dealing with VCs, negotiation, strategy from initial state etc. Of course those are just guidelines, but I was thinking it's good to have all things on the radar instead of trying to invent all business processes myself alongside all the tech that needs to be developed.

Think about whether you need to get an MBA, or hire one.

Bill Gates and Paul Allen hired one* (Steve Ballmer)

* Ballmer dropped out of business school before he finished

If you have the time and money to spend (or it's otherwise paid-for) and a spot at a top 5 program--why not?

Opportunity cost

Thus "_time_ and money to spend." Studying for a year or two at Stanford would probably come out on top of most opportunity cost comparisons, absent a pretty compelling plan in some other direction.

what do you expect to learn specificly?

Not only might there be issues with choosing CEOs based on pedigree and past performance (related to share value), but even the measurement of share value is arguably a weak indicator of company health and profitability.

Many institutions and media organizations are focused on quarterly or annual share price, but those numbers can be temporarily inflated if the company is willing to take actions that would damage itself longer term. And when I say "would", I actually mean "usually do". Also, global and industry factors can push a whole sector up or down for months or years.

I would argue the efficiency of the market produces an excellent correlation between company health and stock valuation over long periods.

But even if you disagree, isn't that just another way of arriving to the authors same conclusion?

If share price is a terrible proxy for company health, then why are we paying CEOs based on it? And if it is a good proxy for company performance, then it's proving that CEOs are being grossly overpaid for what are essentially random outcomes. Under both scenarios, something is terribly wrong.

Just to throw yet another caveat into an already delicious stew, stock price is an imperfect proxy for what we would really be trying to measure, which is the creation of value.

> Headhunters and corporate boards often look for CEOs with a track record of creating value at another company when choosing whom to hire. But if past performance doesn’t predict future results, then they might be looking at an irrelevant variable.

I don't understand this. I would imagine there is lot of accumulated knowledge, connections, network, skills etc. that are involved in taking day to day or strategic decisions.

Not irrelevant at all if my money was at stake.

I just want to point out to those who are excited about confirming their preconceived bias that MBAs don’t matter: according to this result, MBA CEOs outperformed non MBA CEOs by 0.04% per month. That’s quite a bit of alpha.

EDIT: yes, the individual means are not statistically significant but given the large effect

a) where is the p-value b) there is no statistical significance reported of the difference between MBA and non-MBA performance, only of individual means

Except... the study also points out that this is not a statistically significant difference. How do you miss that?

I didn’t. Statistical significance is complicated. In a noisy data set, you have to also look at the magnitude of the effect - and its huge here. Note that they don’t report p-value magnitude, which matters. They also don’t report statistical significance of difference, only of individual effects.

Here's my question: how do you create incentive structures towards sustainable operation? Is this even possible or feasible inside of a company, or is it the market's job to operate such structures, using corporate formation, dissolution, aggregation and division?

It seems very hard to create an incentive structure that won't be gamed.

Author is ignoring factors which are inconvenient due to the difficulty of measuring them.

Let's take a famous example - Steve Jobs. Booted from his own company, then NeXT was a middling success (because of its acquisition), and Pixar was a much larger success but much more because of John Lasseter than anything Jobs did. Was there anything there to suggest that Jobs was going to be as successful as he was with turn-of-the-century Apple? How could somebody who was that successful at the helm not have similarly suceeded in his previous ventures? At the same time, can any sane person say that Apple was destined to take off like a rocket, that Jobs was merely lucky to be in the CEO role when it happened and take all the credit, that therefore anybody else could've been CEO and Apple still would've turned out to be wildly successful? That is patently ridiculous.

People are not islands. People can be more than the sum of their circles. Jobs was as successful as he was because he was in the right company plus at the right time plus with the right people around him (Jony Ive present, John Sculley absent, etc.).

Now, the author doesn't mention this because the author has no way of quantifying the, for lack of a better encompassing term, "cultural fit" of any given CEO candidate with any given corporate setting. But just because the author does not understand the relationship does not mean that the relationship does not exist. That the economy fails to judge CEOs properly and offer the correct opportunities to the correct candidates does not mean that all candidates are appropriate in all roles with their performance more or less in the hands of fate. Getting a random result for non-random inputs is not evidence of randomness but evidence of misunderstanding the inputs or of a poor metric.

Postscript: boards are completely justified in firing CEOs for poor performance ostensibly due to external, macroeconomic factors. The captain's job is to weather storms, not return to Spain with naught but an excuse for why the King's colonial riches have sunk to the bottom of the Atlantic. Maybe the captain hired the wrong sailors who would've fared better in fairer weather. Maybe the captain was brought aboard a boat with a hole in the hull. But part of the captain's job is to make sure the ship is properly staffed and that holes are plugged. If they're not, then the captain doesn't get to cry foul for a storm exposing his incompetence. Jobs had his storms - the dot-com bust and the 2008 recession. So what?

Seems like they could just do a 15 year annuity that did X times the market avg price every two weeks. Make it transferable so CEOs could liquidate some for higher up front salary.

I believe, just my belief, that a "good CEO" can install a great spirit in the company, which trickles down through management to all employees.

I love to hate on MBA types just as much as the next engineer, but this article contains a distracting logical fallacy right away:

That CEO pay increased because of a trend in aligning incentives (the analysis of the CEOs performance says nothing about this).

The article spends no time on this and is at best an ancillary point, it should have started with a general statement of how crazy CEO pay is and if it’s worth it without dealing with the incentive nonsense at the beginning.

Otherwise good article.

But this was the justification in countless think pieces in the Economist, the NYT, and so on. Align the incentives of the CEO with the shareholders and magic will occur. The author is dueling with a ghost that you never saw.

The rise in CEO pay, and tying CEO pay to stock performance, are entirely orthogonal issues. The fact that the article doesn't seem to realize this distinction, makes me question how much of the rest of the article is similarly flawed.

Can you elaborate on why they’re “entirely orthogonal?” Since this increased compensation is stock, aren’t they at least somewhat related?

You can pay someone the exact same target-compensation, while still shifting it away from cash, and towards stock. For example, giving someone a stock grant of 1000 shares of GOOG, instead of paying them $1.1M in cash.

There are still issues with this approach, although they're a bit more subtle. Think about Valeant circa 2015 -- definitely taking a short term approach of acquiring and stripping pharma companies, which did positively impact their stock price. But, they did it in an unsustainable manner.

Does anybody know what Figure 2a and 2b are trying to show ?

Don't know why they included two figures when two numbers would do (it appears the first graph in each is showing the definition of median and quartile).

If the top quartile of performers were consistently the top quartile, then the plots in Fig. 2B would be the same because their performance is persistent over two three-year periods. Same thing can be said of Fig. 2A but with the median.

The simpler and more intuitive way to show this would be a some way to display "skill performance", "chance performance", and "observed performance".

| -------------- | Skilled | Chance | Observed |


| Median------ | 50% | 25% | 25% |

| Top quartile | 25% | 6.25% | 7% |

Applications are open for YC Winter 2022

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact