Are CEOs paid their value added?

Remember Paul Krugman’s forays into “the wage reflects what the top earners are really worth” topic, and the surrounding debates?  Why should this discussion be such a fact-free zone?  Why so little discussion of tax incidence?

Let’s start with the literature.

Read this paper by Kevin Murphy (pdf), especially pp.33-38.  Admittedly the paper is from 1999 and it won’t pick up the more recent problems with the financial sector.  But most of the data are from plain, ol’ garden variety CEOs.  In many of the estimations we see CEOs picking up less than one percent of the value they create for the firm, and all of the estimates of their value capture are impressively small, albeit rising over time.  Never is the percentage of value capture anything close to one hundred percent.  “One percent value capture” is an entirely plausible belief.

You might think this sounds whacky but it makes theoretical sense.  For instance often CEO performance is motivated by equity and options, but few CEOs are wealthy enough to own more than a very small chunk of the company (risk-aversion may be a factor too), and that will mean their pay won’t reflect value created at the margin.  It’s a standard result of agency theory, stemming from first principles.

Maybe you’re suspicious of this work but the way these estimates are done is quite straightforward, and results of this kind have not been overturned.  You can formulate a “pay isn’t closely enough linked to performance” critique from these investigations, but not a “they’re paid as much as they contribute” conclusion or anything close to that.  (And, if it matters, the “conservative” and also WSJ Op-Ed page view has embraced these results for almost two decades, at least since the original Jensen-Murphy JPE piece; Krugman identified the conservative position with the Clarkian perfect competition w = mp stance but that is incorrect.)

You might be thinking “Ha!  Burn on Krugman!,” but not so fast.  Like Wagner’s music, Krugman’s position here is “better than it sounds,” though not nearly as strong as Krugman would like it to be.

Let’s turn to taxation of the top 0.1 percent, and focus on these CEOs.  If the tax rate on their income/K gains goes up, the firm will compensate by giving them more equity/options, to keep them working hard.  In other words, the tax rate on the top earners can be hiked without much effect on CEO effort because there is an offset internal to the firm.  At some margin the firm’s shareholders will be reluctant to chop off more equity/options to the CEO, but the marginal value created by maintaining the incentive seems to be very high, for reasons presented above, and so the net CEO incentives will be maintained, even in light of new and higher taxes on CEO earnings.

But here’s the problem, if that’s the right word.  The incidence of that tax is going to fall on shareholders in general and thus on capital in general.  These top CEOs could even get off scot-free, if the shareholders up the equity/options participation of the CEO to offset completely the effects of the new and higher tax rate.  This is also relevant to the Piketty-Saez-Stantcheva analysis that everyone has been talking about; they don’t see these mechanisms with sufficient clarity.

Moral of the story: it’s harder to tax the top earners than you think.

The second moral is that tax incidence remains a neglected topic, even among top economists.

The third moral is that too many people, including both Krugman and his critics on this point, have been neglecting the literature.

By the way, other assumptions can be made and other results generated, but I am focusing on one of the core cases.


How do you determine the value a CEO adds to a company?

I have the same question. The answer is buried somewhere in that long article.

Yes, it's a crucial point needed to substantiate the thesis, and yet they try to make it look like it's a solid one, even as they make questionable assumptions.

For one thing, do they account for the *risk* that a CEO won't add that value? If 99 CEOs tank their companies, and one makes its market cap increase $10 billion while being paid $100 million, that doesn't mean "CEOs only get paid 1% of what they're worth".

The paper appears to attribute 100% of the change in shareholder value to the CEO, a position that clearly makes no sense. With that assumption, however, it's easy to conclude that CEOs are"picking up less than one percent of the value they create for the firm"

That's a preposterous assumption. I hope you are wrong.

Take a look at the paper, especially pages 33-38, as Tyler suggested. I only browsed, but that's what it looked like to me. Happy to be shown wrong, as it is a preposterous assumption.

foosion is right, the relevant text is at the end of page 31. The other measure of performance they consider is shareholder rate of return, which has the same problem. To be fair to the authors, their interest seems to be CEO incentives and agency problems, for which this assumption is reasonable. That is to say, this paper does even not discuss value added by a CEO and to bring it up as an argument for low value captured by CEOs might even be counterproductive, since value added by the CEO is by definition no greater than the total change in shareholder value. It is interesting to consider ways of estimating the value actually added by a CEO — examine companies where the CEO dies in an accident or is removed through similar force majeure?

Murphy's research in the linked paper is to estimate how large a share of firm value added the CEO gets in his pocket (equation (4) on page 31). How large a share CEOs pick up of the value they create for the firm is simply not addressed in Murphy's paper and cannot be deduced from it either. Thus, there is nothing to contradict Krugman's assertion in Murphy's paper (which of course doesn't imply that Krugman is correct).

I agree, Tyler is mischaracterizing the article. I already knew that CEOs don't own the entire company. You really what to know whether increased compensation leads to better performance.

Indeed it is preposterous. Many employees add value well in excess of their pay. Why let the CEO claim credit for that, as well as for random market movements that he makes no contribution to? No reason.

But there are many who are willing to twist the data into knots tp prove CEO pay is justified, because to admit otherwise is a shock to their world view.

Consider Tyler's comment:

If the tax rate on their income/K gains goes up, the firm will compensate by giving them more equity/options, to keep them working hard. In other words, the tax rate on the top earners can be hiked without much effect on CEO effort because there is an offset internal to the firm.

The assumption is that the pay is earned, not awarded by a Board of Cronies.

I figured. Us CEO skeptics might wonder how much "value" is staying out of the way.

Well in that case, since around the time the paper was written, the average CEO has been worthless, since the total return of the S&P since March 2001 is negative.

Foosion makes precisely the point I intended to make, and it's also the point I expected to find in the paper. This is a fundamental attribution error, where the leader (the CEO, in this case) gets all results attributed to them regardless of whether they had any role in them. It's one of the legacies of primate social structure, and it's pervasive across all human society, and we need to be conscious of this bias in order to bring our views back into line with reality.

Assigning all change in the performance of the company to the CEO is the fundamental error here, not the measurement of the company's performance or the strength of the correlation of the CEO's pay to the company's performance, which are the subjects of the paper at hand. It's this error that Krugman et al are attempting to discuss, and that's the point being missed.

As someone with 25 years of corporate engineering experience, at levels from junior to VP of Engineering to external consultant, with companies large and small, I have seen a consistent pattern. The best CEO is the one who does the least to mess things up, and that's truer and truer the larger the company. The key differentiator between successful and unsuccessful companies in a field is almost always execution, and usually execution by some unsung key people in R&D or Manufacturing, not by senior management. CEO policies that reward good execution tend to make companies more successful, but the true percentage of the corporate performance that the CEO is responsible for is fairly small. He's mostly responsible for getting the resources needed by the people who really do make the difference, and not getting in the way.

Tyler has mis-interpreted it. There paper attempts to show how CEO pay is responsive to changes in shareholder value (because of incentive programs), not the other way around. The results in that paper are consistent with paying CEOs for pure noise in stock prices. p 41-44 on incentives predicting returns is closer (since changes in incentives should only help insofar as they reflect changes in CEO added value), but the author notes that the evidence is weak and inconclusive.

By how great a percentage of its workers he terminates to inflate profits?

How about the number of contracts he breaks and government dictated cram downs (bankruptcy court orders) and government bailouts forced (unwilling pension fund takeovers by the government), and number of shareholders wiped out, as has occurred in the airline industry?

From Forbes CEO 2011 compensation list:

American Airlines:
CEO Total Compensation
$1.225 mil
5-Year Compensation
$14.34 mil
Gerard J Arpey has been CEO of AMR (AMR) for 8 years. Mr. Arpey has been with the company for 29 years. The 52 year old executive ranks 10 within Transportation

Total Compensation
$1.095 mil
5-Year Compensation
$5.97 mil
Gary C Kelly has been CEO of Southwest Airlines (LUV) for 7 years. Mr. Kelly has been with the company for 25 years. The 56 year old executive ranks 11 within Transportation

CEO Arpey's management failure is he refused to take AMR into bankruptcy. One must conclude CEO Kelly's problem is he has failed to take Southwest to the brink of bankruptcy.

Delta Air
Total Compensation
$4.745 mil
5-Year Compensation
$7.763 mil
Richard H Anderson has been CEO of Delta Air Lines (DAL) for 4 years. Mr. Anderson has been with the company for 4 years. The 55 year old executive ranks 7 within Transportation

CEO Anderson gets the big bucks for leading an airline that failed and broke its contracts and promises, but the winner is, before giving up CEO in 2010:

$5.975 mil

$22.30 mil

Glenn F Tilton* has been CEO of UAL (UAUA) for 8 years. Mr. Tilton has been with the company for 8 years. The 62 year old executive ranks 6 within Transportation

CEO Tilton oversaw one of the largest and lengthiest bankruptcies in US history.

The US passenger airline industry is one where no one seems to have figured out how to remain profitable without government regulation, or treating airline passengers like hauling cattle, as in the case of Southwest. The CEO of of the "air travel is a no frills commodity" carrier gets paid the least, while the CEOs of the airlines that try to sell luxury, convenience, and service get paid a lot for failing to deliver, and then going into bankruptcy for another do-over.

Why would CEO pay go up if tax rates go up? Isn't it set by supply and demand? Is there any evidence that ceo supply would go down steeply if tax rates went up? That people with these skills would simply do no paid work in sufficient quantity? Possible, but I would like to see evidence.

Boards of directors often look at CEO pay on a net basis. The clearest example of this is excise taxes paid on change of control payments (golden parachutes). Companies often indemnify CEOs for these taxes.

I'd like to see some evidence CEO pay is set by supply and demand. I mainly see evidence that boards believe their CEOs are above average and should be paid above average compensation. Recall that not everyone can be paid above average, but the effects of trying leads to an upward spiral.

@foosion: "I mainly see evidence that boards believe their CEOs are above average and should be paid above average compensation."

foosion, that *is* a demand effect! (To see that a price to be set by supply and demand, one needn't believe that the demand is rational! Astrologers' wages are set by supply and demand, and pointing out "Astrology doesn't work!" is not a counter-argument to that proposition.)

@Gene: that's not the normal supply and demand mechanism

Gene_Callahan, it is very ignorant of you use astrology as an example of something that "doesn't work". You're thinking about the sort of "man in the street" astrology, not the kind practiced by serious, top astrologers. And astrology is about spirituality, anyway.

tongue in cheek?

CEO pay is set at aspirational levels in the same way that luxury cars or college educations are. If a board wants an above average CEO, they feel it is necessary to provide above average pay. Because every board wants an above average CEO, every board feels the need to pay its CEO above the average. And so goes the one-way ratchet of executive compensation forever upward.

Why is this particularly true of CEOs? doesn't everyone want an above-average everything? why aren't salaries for engineers, doctors, lawyers (well, okay), teachers, pie-bakers all similarly ratcheted?

As an engineer, my salary is set by a manager who is compensated partially based on his ability to contain costs (keep my salary realistic) while getting the job done (keep me from leaving). Nobody writes a contract for engineers that has a ratchet on it that I have seen, and it would be tough to get accurate comps that would be legally enforceable.

Literally none of that applies to CEO's. Their salary is set by people who have an interest in seeing CEO pay ratchet upwards. They write contracts that explicitly tie the pay to the pay of the comps (the number I saw somewhere is that 75% of CEO's have that clause in that contract). And if they have a hard time finding comps, they just comp them to the largest and best companies that are remotely in the same field. Since the number of comps is less and is publicly available, you can write a binding contract (CEO will be paid x millions or 10% more then the average of acme, bcme, and ccme CEO pay, whichever is more).

Thank you for the clear answer!

@Babar: "Isn’t it set by supply and demand?"

By raising taxes on the product (the product being "CEO work") you have shifted the supply curve leftward. If the demand curve is sufficiently steep, the result will be almost entirely an increase in price paid rather than a decrease in quantity supplied.

Richard Posner's response:

Notice that both the size and the value theories imply that the supply of highly skilled managers is inelastic. Were it elastic, the increased demand for managers would not result in higher pay, but simply in an influx of qualified persons from other activities (including the management of divisions of large firms). The inelasticity of supply of highly skilled managers implies that the higher pay of these managers is a scarcity rent, which could be taxed away without reducing the supply significantly, though there would be some reduction because the elasticity of supply is not zero.

KLO, are you Richard Posner? If not, why are you posting something called "Richard Posner's response" without sourcing it?

It'd have taken you less time to google it than to type out that gratuitous rebuke to KLO.

Sorry. I assume that the source can be easily found using google. I was copying it from a hard copy, so I did not have the link.

i'm not denying that there could be some increase to the price paid, but if taxes went up on CEOs what leverage would they have to get more compensation? they could walk, for sure, and not work - but how many would?

I would love to see them go Galt.

If one compares the incomes of top earning entertainers with those of Fortune 500 CEOs, the entertainers come out on top by a wide margin. Moreover, the entertainers tend to be uneducated and earn their money earlier in life while CEOs go through many years of education and 80 hour weeks in middle management before they get the really big bucks. Compare the Forbes Celebrity 100 list with their top paid CEO list.

From either a moral perspective or the perspective of steering the young onto a career path that is beneficial for society, a good case can be made that EVERY Fortune 500 CEO should make more than ANY entertainer.

The CEO has a major impact on the well being of tens or hundreds of thousands of employees and shareholders. The entertainer makes their pay only because modern technology has made possible a winner take all environment in entertainment. Had Lady Gaga been born 100 or so years earlier she would have made a better living as a member of the oldest profession than as a pop entertainer.

The CEO career path is already far more attractive financially than the entertainer path. Very few people actually become a star or a CEO, but the lower tiers of business management are also very lucrative, while most of those who fail to make it big in entertainment cannot make a living in the industry at all.

If we were to take Robert Frank seriously, that would be another reason for higher taxes on entertainers.

If an entertainer stops entertaining, they stop getting paid. If a CEO slashes American jobs and transfers sensitive tech to China, they get a raise. If they crush a company and cause mass layoffs (cf HP) they get a fatty golden parachute. At least Gaga didn't fire anybody because she was incompetent. A good CEO is more important then an entertainer - and most vastly outearn the median entertainer over a lifetime by a very very wide margin. The problem is that too many CEO's are worse then baboons, they actively harm their companies. AND THEY STILL GET PAID. Also, the only "entertainers" on this list are Spielberg and Lucas, and I'm going to guess it's more from their merchandising empires then from the actual movies.

This is, to a large degree, result of the deformation of the market by excessive copyright protection (a typical governmental privilege).

The only reason why Hollywood can afford paying billions for movie stars is that Uncle Sam will prosecute copyright breakers for taxpayer's money. If the enforcement wasn't there, the wages of most ''stars" would drop to normal middle class brackets.

Hollywood pays for stars because stars sell tickets.

Flamebait is flamebait. Though I do want to point out that Lady Gaga is from a relatively well-off family and 100 years ago could have been a Senator's wife with just as much influence as she has now.

"Maybe you’re suspicious of work coming out of U. Chicago"

In which case, you should stop being a partisan dope.

See also the report by the High Pay Commission in the UK (it has an Exec summary):

And even accepting the notion highly-paid CEOs, lawyers, professional athletes (why are they usually ignored?) etc don't actually produce economic benefits commensurate with their pay, that doesn't seem like a great rationale for higher tax rates. These outsize paychecks arise from either bad legal frameworks and should be addressed therein (caps on legal fees, stronger shareholder rights relative to the BOD), or are the result of the often capricious nature of utility.

If there are features of our society, culture, and economy that lead to incomes that have no link to economic value, and those features are so widespread and deeply entwined, then why wouldn't taxation be a preferable way to address such externalities?

If the choice is "change the entire corporate governance structure of US corporations" or "raise the marginal tax rate on people in the top 0.1% income bracket back to the rates of the Clinton era", the more conservative choice - and the only plausible one - is the latter, not the former.

But who decides that someone's job has "no link to economic value"? Robert Frank? I don't think I'm down with that.

That's exactly why we have elected officials. Don't trust them? Get involved.

Elected officials should get to decide who is worth what, really? Yikes.

CEOs should not be paid based on some nebulous estimate of value added to the firm, or even stock performance, they should be paid based on their replacement cost like everyone else. There is no immediate replacement for Lady Gaga at the moment. The Packers don't have another Aaron Rodgers. I would argue that there exists an immediate replacement for Jeffrey Immelt. That doesn't mean his compensation won't still be incredibly high. Boards of large corporations have favored executives over shareholders and in my opinion the rules of shareholder elections and the peverse demand for "independent" board members exacerbate the problem.

I would argue that there exists an immediate replacement for Jeffrey Immelt.

These folks come immediately to mind: Barney Frank, Barack Obama, Maxine Waters, Mitt Romney, Rick Perry, George Bush, etc., etc., ad infinitum, ad nauseam.

Or anyone at the C-level at GE and/or any other Fortune 50 company.

"There is no immediate replacement for Lady Gaga at the moment. The Packers don’t have another Aaron Rodgers."

I kinda disagree with this. If Lady Gaga didn't exist, people wouldn't stop buying records (mp3s), they'd just be buying someone else's mp3. If Matt Flynn were quarterbacking the Packers they'd still have a season ticket waiting list 30,000 names long.

If anything the only gain is in consumer surplus - the people who are buying Lady Gaga mp3s apparently get more than 99 cents' worth of enjoyment out of her single, say $1.29 worth. If they were buying an mp3 of an "inferior" artist perhaps they'd only be getting $1.19 worth of enjoyment. Similarly, Packers fans would slightly less enjoy watching a Matt Flynn-quarterbacked team, because presumably they'd lose more (though the NFL is largely a zero-sum game, so the loss of enjoyment of the Packers fans would presumably be offset by increased enjoyment of the fans watching their teams beat the Packers).

Perhaps you're overthinking the comparison. Without Lady Gaga, all Lady Gaga related earnings vanish. Lady Gaga can't hand the act over to her brother like Gallagher and expect it to work. Without Aaron Rodgers the Packers are not 11-0. Without Immelt maybe GE still has its AAA credit rating and its stock isn't down 50%. Maybe thousands of other executives could achieve similar results at GE. That fact should be driving his compensation.

And anon, I've tried and tried but I don't understand your comment.

David - anon is contending that Immelt's value to GM is the value of his political connections and the ability to get government contracts/bailouts/etc.

+1 to David N.

The delta between Immelt (or any 'top' CEO) and the next best available person is pretty small.

Same goes for Wall Street by the way. If John Paulson has a fatal stroke tomorrow, does society's utility decline at all? Maybe his fund has a little lower return, and some other fund makes more...why does John Paulson make so much money?

...especially now that his fund is down 45%.

"There is no immediate replacement for Lady Gaga at the moment."

You seriously underestimate the share of reasonably good singers per population. Outside the glamorous world, there is a deep, murky ocean of good musicians who are never fished from obscurity by the Big Recording Industry, and play around the bars.

No I don't. Lady Gaga is apparently a poor choice as an example because it brings forth so many latent biases in people towards pop success. Her phenomenon is not about singing ability.

Artists, authors, etc. are not employees of the companies they have distribution deals with. Their absolute compensation is not negotatied or limited by those companies. In the context of "Lady Gaga Enterprises," there is no substitute for Lady Gaga. Beatlemania did not earn at the same rate as the Beatles. Jeff Immelt is clearly a talented executive, but in the context of GE there are many ready replacements and his compensation should be a modest premium over the next choice.

You never want to pay someone the value they add. Profit is the difference between revenue and expenses. CEO salary is an expense.

Which is why, if we want to stay competitive, we need to offshore CEOs. You can get 5 CEOs in India for half the cost of a US CEO.

That's just makes good business sense. What do you have against globalization anyway? Are you racist?

we need to offshore CEOs

We are gradually moving up the value chain for outsourcing and offshoring (e.g., see legal research and medical tourism).

Give it time, give it time....

Happening already. Land Rover and Tetley both were acquired by Tata Group. Most of the deal financing was external.

What Tata's really provided was cheap top management.

LOL. I predict in a few more years the NY Times will be clearing her throat and making polite noises about "the American way of doing business" and the good old days when filthy rich executives spent shareholder money on things like the ballet and public broadcasting.

Working class whites and blacks marginalized by mass Central American immigration ... no problemo. Using immigration policy to increase the supply of upper management and white collar professionals ... this cannot stand!

>Moral of the story: it’s harder to tax the top earners than you think.

No, it's just harder than Krugman thinks. Er, "thinks."

Rich people have plenty of tax-avoidance options, especially given the political class we are stuck with. Sane, rational people have understood this for a very long time. Useless partisans hacks still struggle.

Many (most?) top earners are not CEOs, but rather financial professionals, elite services providers (e.g. tippety-top lawyers), entertainers, or entrepreneurs

I eagerly await all those CEOs whose companies' stock price has gone down to give back 1% of the value they destroyed.

Resentment of high pay is not resentment of success. It's the failures that rankle. Think Jon Corzine is hurting as bad as his employees? Do Oakland fans feel good about how Jamarcus Russell turned out?

That's the risk they take. OTOH, a CEO or athlete who does well can be worth much more than they make.

Yeah there's not a lot of resentment towards Jobs, Gates, or Buffet. It's the out of control "Lake Wobegon" effect that is the problem. All these weasels are getting paid like they are above average when many are not only below average, but active destroyers of shareholder value. Nardelli getting 200mil in severance for treading water at home depot (DURING A HUGE HOUSING BOOM... jebus it should have been like shooting fish in a barrel... Lowes stock trippled over that period), or Apotheker getting 26mil for a year of slashing shareholder value at HP by 40%.

"Maybe you’re suspicious of work coming out of U. Chicago"

If only this was the Kevin Murphy from the U. Chicago. Kevin J. Murphy is the expert on exec compensation at USC. Kevin M. Murphy is at Chicago.

I'm sensing a new Murphy's law...

Though the two were, I believe, classmates at Chicago and earlier at UCLA.

CEOs, investment bankers and asset managers, some lawyers, and a variety of other high earners, all differ from more ordinary income earners in one key respect. They are paid not per hour or day or month, but essentially on a commission basis: colloquially called "a piece of the ups". As a function of assets under management, a jury award, an increase in equity value, etc. A stock option is different only in form from a broker's commission, or a hedge fund manager's fee. Entertainers often get " a piece of the gross". Even in the NBA labor pact there is a commission for workers (50% of "basketball-related income"). And in a world of fiat currencies and declining interest rates, the size of what they are basing the commission on increased vastly, while others whose compensation metric is based on time periods, don't, because the hour, the day, the year don't grow longer.

Like wages, these commissions generally only go one way though, there are no negative commissions. This is different from what investors face, as they can incur not only profits but also losses on investments, which risk may justify a lower tax rate on investment profits.

In sum, there are three kinds of compensation - wages and salaries based on time periods, which are relatively stagnant but have no risk of loss; upside commissions tied to the size of something the earner is responsible for, the size of which has increased dramatically, I think due to fiat currencies and declining interest rates, which also have no risk of downside loss; and profits on investment with concomitant risk of loss.

This is a very good post.

It seems to me that the first part of your post is a failure in counter-factual thinking. The question should not be how does CEO pay compare to company growth, because that attributes 100% of the growth to the CEO, which is silly. The question should be, if I payed offered my CEO position at $1 less, how much would that effect future growth in stock value. There 1% seems laughingly small, as a $100k change in pay per year would have to result in an average $10milllion growth every year in stock value.

Statistical analysis of baseball performanence often talks about a player's contribution relative to an easily available replacement. For example, major league pitchers are evaluated relative to what a pitcher at the highest levels of the minor leagues would do. Instead of assigning 100% of the increase in shareholder value to a CEO, we should compare how a company would perform with that CEO to how that company would perform when run by a VP or some other internal promotion. The marginal difference in salary should be related to the marginal difference in performance. Comparing salaries of low-cost executive promoted from within, to salaries of big-name hired guns, and comparing performance, you can easily show that executive pay is not rational from a shareholder's point of view.

Instead of executive pay being rational for shareholders, it is rational for corporate board members. Since a seat on a board is so lucrative (five to six figures per day), a rational board member's first priority is to not let the company go bankrupt, and the member cares more about limiting downside risk than creating marginal value to the company shareholders. Limiting downside risk means never (or very rarely) hiring an inexperienced CEO, which artificially restricts supply, and increases prices. Corporate boards are OK with this increase in costs, since they view it like an insurance plan and they benefit disproportionately from the reduction of risk. Shareholders and the rest of the economy should not be OK with this.

Minor correction: a board member's most immediate concern is remaining on the board. If you take a risk by hiring an unproven, cheap, CEO and he fails, you might get ousted. But if a well known, expensive, CEO fails, you can cya and say "well we hired the best - the circumstances were out of our control."

You also see this in the legal arena. Companies hire extraordinarily expensive white shoe firms over just about equally competent, cheaper firms, because the high price alone is some sort of signalling mechanism that says We Are Taking This Seriously.

Good post. Now somebody needs to get cracking on a SVOR metric (Shareholder value over replacement).

Good idea, although I think the reference to VPs is not correct, either as analogy to baseball or on its own. The right comparison is to other CEOs in the same sector. In baseball, we don't compare starting shortstops in the majors to minor leaguers or bench players. We compare starting shortstops to other starting shortstops. So here CEOs should be compared to CEOs in the same sector. Proxy statements do provide some graphs on that.

one of the most interesting recent (2008) paper on the topic is Landier and Gabaix :

It answers questions on the link between firm size and CEO pay.

To summarize much of the commentary for Tyler: The reason people don't talk much about the data when discussing CEO pay is that all the studies seem to be laughably bad. Until researchers can begin swapping CEOs around companies in pre-determined patterns in double blind studies — so, until hell freezes over — there won't be anything like proof positive for either side. That's not to say we can't theoretically have studies that tell us something, but any real-world study I've ever begun to read has had such obvious and massive problems that I wonder how it got published.

This is an example of spurious economics.

Ask yourself this question: what accounts for the difference in CEO salaries over time? Were the managers of one period dullards and the managers of another periods suddenly geniuses.

Now, if you ask this question, you would have to say: if CEO salaries as a percentage of gains were lower in the 1950's, 60's and 70's and higher in the mid 1980's forward, what accounts for this> This is where the gains have been made. And, since CEO comp as a percentage of profit was low in those earlier periods, you would have to ignore the major growth of productivity in the 50's through the mid 70s.

So, what accounts for the time series difference?

Changes in corporate governance laws: as a backlash to greenmail activities in the 80's, legislatures and lawyers began drafting tighter corporate control mechnanisms--so there would be fewer tender offers, greenmail, etc. The unintended effect was to put more control in the board, more control the the executive, and less in the shareholder.

Surprise, surprise: a corporation now protected from takeover threats rewards its management differently than one that isn't because the manager now if it selects cronies on the board doesn't have to fear a hostile takeover.

Where have you gone, Gordon Gekko?

Nice post by the way.

"There is no immediate replacement for Lady Gaga at the moment. "

I disagree. Lady Gaga herself is a replacement for Madonna, once she got too old for the role.

It seems popular culture currently demands one successful entertainer who who carriers off the Madonna/ Lady Gaga persona. There are a number of young female singers who can fill that role (and yes, you have to have a certain level of talent), but only one gets to be Madonna or Lady Gaga or whoever the next one will be.

I think this was Taleb's point. Someone is always going to be in one of the long tails of the bell curve. Its not like the long tail goes away if you remove whoever is currently occupying it. And I think he was referring to CEOs as well.

This is what I was trying to say. For a more recent example, one multiplatinum selling British born female neo-soul singer dies (Amy Winehouse), and all of a sudden we have a new multiplatinum selling British born female neo-soul singer (Adele).

See my 2nd comment above. Also, Gaga makes the tail long, she doesn't merely occupy an existing tail. The demand for pop music is not zero-sum.

Gaga's replacement doesn't have to be in pop music. The tail is defined by the source of its revenue, not by who occupies it. Young screaming girls will find another idol on whom to shower their parents' money, because that's the nature of young girls, and that's the nature of our culture.

The problem is that while a replacement would surface there is no guarantee that that replacement would sign with the same record label. Some other tart could become the next big thing but you aren't making any money off it, some other people are. So if your a part of the Lady Gaga gravy train you don't want to play the lottery again trying to find a new Lady Gaga, not until you have too.

Yeah but that's not a problem for society at large, which is generally indifferent as to which record company its pop starts sign with.

Whether its a problem for society at large or not doesn't effect the value of Lady Gaga to her publisher. And they pay her based on her value to them, not her value to society.

You could make the same argument for CEOs. Society doesn't really care if the CEO of Coke crushes the CEO of Pepsi with witty ads or vice versa. We will find our sugar water. So the marginal value of both of those cats is zero, and they should be paid as such.

Its not like the long tail goes away if you remove whoever is currently occupying it

Specious. We can't just assume that if all today's popular entertainers had become engineers instead, the market for entertainment would not be smaller.

The market might be smaller as measured by dollars spent on that market, but that's sort of the point. Excessive compensation of entertainers leads to too many resources devoted to entertainment - and not a lot of extra utility for the society.

We can’t just assume that if all today’s popular entertainers had become engineers instead, the market for entertainment would not be smaller.

You can't assume that it would be, either. Everyone knows indie acts that are as good as or better than the big-name, big-dollar, heavily promoted ones. Clearly the top entertainment positions are not going by merit, so why would you expect reshuffling them to someone else to produce any systematic effect in any particular direction?

There's a framing problem. Who pays Lady Gaga? Her fans. WIth many goods, not just pop songs and movies, the major competition is not another store, singer, or film, but no sale at all. You can't frame the issue as if Gaga's failure to sell a ticket guarantees those funds to another perfomer. More likely it's Gaga or stay home.

Who pays Jeff Immelt? GE shareholders. They don't have a "no CEO" option. What they do have and many other companies have is a board that negotiates poorly with management, eschews long term thinking, and is almost impossible to replace.

It's actually not GE shareholders who pay Jeff Immelt. The shareholders have no money to hand to Immelt until they get it from customers of GE. The shareholders simply decide how to divvy up the money they get from the customers. Actually, it's not even the shareholders who do that, it's a board of directors that supposedly represents the interests of shareholders. They divvy up the revenue. Some for Immelt, some for labor, some for shareholders. Ultimately GE's customers pay Jeff Immelt, and GE's workers, and GE's shareholders. You. Me. We pay for the errors in judgment made by a board of directors that allegedly represents the interests of GE shareholders.

It's a distinction without difference. What you describe is true of all employees of GE, not just Immelt.

Andrew, you seem to be willfully ignoring my point. Gaga is in the tail not because she has the "young screaming girls." She's in the tail because she also has Greg Mankiw. There have been a lot of new #1's since Elvis and the Beatles, but no bigger sellers (except for Zamfir of course). Why does Heinz have 60% of ketchup sales? Is it the nature of our culture?

Also, your logic regarding yourself as an unforced GE customer paying for Immelts's "mistakes" is completely bogus.

The incidence of this tax will fall on the shareholders.


Since foreigners own shares of US corporations, they will be paying my taxes!

Shorter argument: execs didn't *really* get an unfairly good deal--IF you ignore the once-in-a-century crisis that wiped out trillions of dollars of wealth and threw millions of people into unemployment, poverty, and out of their homes, while somehow leaving all those important, "fairly-paid" people in charge in tact, with lots of money, no fines, and no jail time, as they availed themselves of their own personal Fed and congressional channels of lender of last resort, that were of course used to take their "fairly"-earned bonuses, and that were kept out of reach and secret from hoi polloi.

Why, indeed, aren't more Americans buying into Tyler's argument? What is it that they can't get through their thick skulls?

"By the way, other assumptions can be made and other results generated, but I am focusing on one of the core cases."

Sorry, but that sounds like a lawyer-joke punch line, Tyler.

I think the generous assumption would be that you focus on this paper only because it attempts to apply a method Krugman claims to rely on, not because you take it seriously yourself. That would, for instance, explain why you didn't more closely read Murphy's, um, generous assumptions about CEO value.


I wonder how the CEO pay would look like if companies had a little more direct democracy. If every shareholder was allowed to vote on a CEO-package would most CEO's get less or more?

What would be the downsides of such an idea? In the past maybe it was unfeasible; but with today's IT systems more participatory democracy shouldn't be that hard.

Evil arises out of the cesspit called Board of Directors.

Probably more variance. Way less pay when the stock is doing badly and way more pay when its doing well.

Because frankly if Jobs had wanted to get paid more, I don't think any Apple Stock Holder would have cared.

Clearly the value of a firm is not due to the CEO alone as you might suggest, what about other people who work for it ? Are their contribution valueless ? Maybe the opposite is much more likely: bad management can eventually drive a firm to ruin. It is much easier to lose a battle rather than to win one, i.e. if one general wants to lose (provided that his opponent wants to win) he certainly will lose.
And what about the other side of the coin (the offer side) ? How many people would fight to be hired for less than 1% of the increased value of the firm during their 'reign', considering that the standard salary package is made by a relevant fixed sum plus some variable components ? The problem is that CEOs salary are controlled by boards which are not indifferent to their levels. What is worth more for the board: the risk to be fired together with a bad CEO or the spillover effect of generous CEO salaries ?

The idea that a CEO should capture all the corporate value added during his/her term is ludicrous. Unless, of course, they are also on the hook for all of the value lost when things go the other way.

Why do CEO's in the US get paid so much more than their counterparts in other countries, say Japan, for example? Also, it seems quite simple in theory to justify the amount of CEO salaries based on the worth of the corporation, but isn't the current social upheaval going on at present based more on the questions as to whether it is morally & ethically correct for these individuals to be living the Fat Cat life while so many others are unemployed and struggling?

Thats easy: Japanese CEOs are allowed to insider trade.
Technically its illegal, but japanese culture says its absurd to ban it, so its not enforced.

It is extremely difficult, if not impossible, to determine the value that a CEO adds to a company. What is certain is that it is only a small fraction of the increase in the market capitalisation of a company during his/her tenure. For a start, most of the value-added in a company is created by its myriad employees, and at the senior management level, any good CEO relies on his/her team. Second, good decisions take time to have an impact and good management structures often endure, so success that is attributed to a CEO is often due to things that happened before his/her tenure. Third, share prices rise for all sorts of reasons that have nothing to do with the CEO: lower interest rates, faster economic growth, a financial bubble etc. In short, a CEO's impact is vastly overstated - and his/her success is often due to luck not talent.

Is Krugman unaware of the concept of consumer surplus? Even in the perfect competition model, you do not capture all of the value you create.

Even if the CEO is worth as much as is proposed (which sounds dubious) I think a valid question is what is the CEO worth above a replacement CEO, i.e. one who could be hired for far less money. Is the CEO in question really that outstanding that he is creating the value or is it his position which allows the creation of value as long as he is reasonably competent?

The value of the S&P 500 is a function of earnings growth and the discount factor for future growth or the PE. Supposedly earnings growth is largely a function of how well the CEOs do and the PE is largely a function of inflation and interest rates.

If you look at S&P 500 earnings per share growth since WW II the trend growth rate has been about 7%. Although there are cyclical swings in earnings around the 7% earnings trend there is no evidence that their has been any shift in trend EPS growth since WW II. This strongly implies that even though shareholders have been paying CEOs much more since about 1980 they have gotten nothing in return for their payments since there has been no change in trend EPS growth since CEO pay shot up.

In other words shareholders have a major agent problem.

This is completely embarrassing. This paper emphatically does not say that CEOs are paid less than one percent of the value that they add; it merely says that they receive less than one percent of increases in shareholder value, because that's the amount of stock they own on average. The paper makes no attempt whatsoever to determine whether and to what extent the CEO was responsible for the increased shareholder value, so it only says what Tyler says it does if one believes (against all notions of plausibility) that 100% of changes in shareholder value are attributable to the CEO.

I remember there was an article somewhere that most CEOs do not provide any additional value compared to the firm hiring any other CEO. In other words, as long as there is a parameter for "CEO material" basically anyone who falls into that category would do the job just as well. Only a very few CEOs are actually "value added" in the sense that they outperform other potential CEOs. Almost all these superior CEOs came out of Jack Welch's GE although I think a few rare outliers like Steve Jobs also exist.

If so, this has lots of implications. The first being that there is little need to increase CEO pay. Most CEOs won't earn the higher pay. Regardless if you pay one CEO high compensation or hire another for less, it won't add much to the bottom line. Any extra pay above the general minimum for a CEO doesn't get you any better performance. Only a few superstar CEOs would actually deserve greater compensation.

As the other posters have stated, you can't attribute increase of shareholder value exclusively to the CEO. In fact, most probably have little to do with it. Some probably decrease shareholder value.

I think these links might be what I am thinking of

If anyone has a better memory of the article/study, please post.

Yeah, this is embarasssing.

Post Hoc Ergo Prompter Hoc? The Market Cap went up ergo the CEO added all that value?


For those of you wondering whether increased pay leads to better performance, I recommend watching this:

"You might be thinking 'Ha! Burn on Krugman!,' ” Well you know, Cow. I'm sorry. I mean Tyler. That's exactly what I was thinking. Here's why:
1. Paul Krugman is not a legitimate econ-o-mist and cannot be trusted. Period.
2. I was CEO of Godfather's Pizza. I added value to Godfather's Pizza by recognizing that a manly man don't want his pizza piled high with vegetables. He wants meat. And we gave him meat. They compensated me less than 10% of the value created from that idea.
3. 9-9-9

There is a cool paper measuring the influence of leaders on country's economic perfomance using random (i.e. not by assasination) deaths of leaders (Olken and Jones I believe). I wonder if enough CEOs die in office to come up with something similar.

Only if in car accidents, etc. With deaths from natural causes, onset of dementia will be dominating factor :-)

Many car accidents result from risk-taking behavior (although not necessarily on the part of the deceased). And of course miles driven play a big role in anyone's risk of car accident, fatal or otherwise.

I wonder if there's any kind of death that is truly random (i.e. not correlated with any characteristics or behavior of the deceased). Probably if there is, it's too rare to form the basis of this kind of study.

I am shocked, yes shocked, to see some of the comments here about CEOs being rent seeking. I didn't realize this. Surely we must pass a law that CEOs must be saints who worry only about the social impacts of their companies on a strictly utilitarian basis. And we will have to get some regulators (also saints) who will ensure this is true.

Meanwhile, in the real world, I don't worry about CEO's pay. If I don't like what a CEO is paid I sell his company. And so the company gets a very direct measure of the CEO's value. So even when they (inevitably) rent seek, society can keep their rent seeking down to tolerable levels. Unless the Government steps in and makes the company a quasi-monopoly due to their company being too big to fail or something.

People, private rent seeking by CEO's is not a major problem, let the owners worry about that. The biggest problem is society right now is the public rent seekers; trillions of dollars for Investment bank bailouts.

here is one reason why rent seeking ceos are a problem
In general, luxury goods have higher margins then nonluxury goods.
so,if you have a small number of people with high incomes, they can distort the entire economy
as a specific example, i live in newton, ma, an affluent suburb just west of boston.
over the last few years, many of the stores that cater to ordinary people - the used clothing store, the low cost diner, etc, have been replaced by stores that cater to the wealthy (personal trainer stores,)
this is because the small % of people who are wealthy can support more stores, with higher rents, then the large % of people who are not wealthy,
as a result, the lives of the majority of people are made harder.

Ezra, no offence, but you haven't shown that the overall utility of the situation you describe is better than the alternative you would want. For instance, the owners and people who work in these shops probably have a better job, which (potentially) offsets any loss of your utility. I don't say this is necessarily true, just that you haven't shown it.

In any event, I am not claiming that rent seeking CEO's don't bother me, but they bother me much less when I have to contribute to their rent whether I like it or not. I'll say it again, but the biggest problem to be addressed is not CEO's who are taking rent from shareholders (who can look after themselves), its rent seekers in the public sector, like investment bankers, who claim to be risking private capital, and are actually risking public (i.e. mine and yours) capital without us being able to do a thing about it. And actually, CEO pay is small beer compared to the money those guys are raking in, its billions versus millions. I don't have a good answer to the problem, but I would rather see the discussion focus on solving that than resolving what is essentially a private matter between shareholders and CEO's.

Executive pay is not millions. Even just S&P 500 and call it 5 mill per is a couple of bill. Add in the other C-execuitves and add in the board (mid 6 figures for a part time job) and you have real money. Also figure in the sums being paid to executives that no longer work there. There have been a bunch of breakdowns of the top .1% and finance is a huge part of it, but nonfinancial executives is also very big. The idea that people who are average at their job deserve these staggering amounts is corrosive to America. Why are Fannie and Freddie such a mess? Because they tried to generate profits that would justify private scale (not government scale) wages for their executives.

I don't think the problem is necessarily rent seeking. I'm assuming most overcompensated CEOs at least try to add value. But I do think some changes to the law and to regulations can help. In my opinion independent boards have had no positive effect on the prevalence of short-term manipulation, protecting shareholder interests, or preventing criminal activity. Look how independent Enron's board was. Or Worldcom's board. A lot of these folks are professional seat fillers, sitting on multiple boards. They're sold a story by management and have no incentive to ask difficult questions. It's a big problem.

Now, what does the fact that it is even possible to publish a paper with a key assumption like that (100% of the change in the company's value = value of the CEO) tells you about economics as occupation? That the great thinker and renown economist hosting this blog can take this assumption at the face value - what does the fact tell you?

Hey, let's extend this inanity further: Obama's value = GDP of the USA!

I don't know how many of you heard Tyler debate Matt Yglesias on NPR this morning; as a liberal who thinks we should have >80% marginal tax on all income (cap gains, etc etc)over about 1e6$, I thought tyler crushed matt; yglesias doesn't understand how to frame an argument for radio, while tyler was clear and convincing

I heard it and thought that Tyler was way off on schooling. From my reading of the evidence USA schools preform as well as any in the world.

IMHO schools in the USA could be much better by teaching different things but not teaching more and better but no country does better in this regard. They could also do what they do much, much cheaper. Those are not the points he made.

"You might think this sounds whacky but it makes theoretical sense.....Maybe you’re suspicious of this work but the way these estimates are done is quite straightforward, and results of this kind have not been overturned. You can formulate a “pay isn’t closely enough linked to performance” critique from these investigations, but not a “they’re paid as much as they contribute” conclusion or anything close to that. "

After RTFA and seeing what a thin reed it is (yes they really do say that 100% of the gain in the valuation of a company is due to a CEO), you sound super defensive. Probably because you know CEO's who could be replaced by baboons to better effect, but if you admitted that then cries to reduce the amount of loot they are hauling out of the US economy would be tough to ignore.

Are workers paid their value added? Probably not.

Are CEOs in aggregate paid less of the value they create relative to their aggregate salary, compared to workers? Probably.

Are CEOs in absolute terms paid less of the value they create, compared to CEOs? Probably.

Do CEOs benefit from the current state of affairs by becoming stockholders themselves? Probably.

Do workers benefit from the current state of affairs by becoming stockholders themselves? Probably not.

Most stock holders, customers and employees do not care much about CEO pay because it is a small part of corporate profits. So my dividend is a few percent lower than with a low paid CEO the percent is not enough to bother with.

Should we take Tyler's lack of follow up as a sign of his embarrassment at this posting? As an engineer, suggesting that it is the CEO who added my value is more than a slap in the face.

Krugman is a complete retard, as always.

Comments for this post are closed