Wow. That's . . . odd.

[xiphias]

Someone on my friends list was discussing how they were upset by the idea of Congress capping CEO salaries. Fair enough. Someone made the comment that "it's like they don't even know why CEO salaries are as high as they are."

I commented that I don't actually understand why, either -- why are they that high? (I understand the historical reasons: as CEO salaries began to skyrocket in the early Eighties, an attempt to shame corporations into getting their salaries back into historical norms was made, by investigating and publishing said salaries. Which had the opposite effect: it became a dicksizing contest to pay one's CEO more than anyone else, and CEO salaries went up by ANOTHER order of magnitude. What I don't understand is the economic reason.)

The person said that it was to attract the best CEOs.

I said that there's no correlation between CEO compensation and CEO performance.

They said that it's hard to quantify CEO performance, because someone could do really brilliantly, but, for circumstances outside their control, still lose billions in a quarter.

I said, okay, there's no demonstrable correlation.

They said, c'mon, not all things that are demonstrable are quantifiable.

I said, look -- Sam's Club, BJ's, and Costco are in the same market segment. Sam's Club is way behind the other two, BJ's is in second place, but Costco is the market leader. Their CEO makes $100,000 a year. And I asked him if he could point to an industry in which the industry leaders had the highest paid CEOs.

He said, and I got this through comment notification, " in short, you're not here to discuss fundamental principles." and defriended me so that I can no longer see or comment on the post.

. . . the weird thing is that, to me, actually looking at real-world data is kind of PART of discussing fundamental principles. I thought that starting from dogma, working outward, and deciding that THAT was reality, without actually CONSIDERING reality went out of fashion when people noticed that Aristotle had been wrong about the number of teeth that women had.

Comments

[sinboy.livejournal.com]

It'll certainly attract the greediest CEOs.

[xiphias.livejournal.com]

Maybe.

Frankly, I have no idea WAHT it will do. As far as I can tell, when you get to the amount of money that the CEOs are making, pretty much everything that we understand about how people are motivated falls apart.

I don't even know if "greed" covers what goes on.

I'd like to see it studied, anyway.

[linenoise.livejournal.com]

Dogma is immune to data. Alas. It's visible on all sides of the political gang war, really. It's also generally the case that the weaker one's position is, relative to real-world data, the more strident one's assertions of "principles" and the idea that we're all just being unfair.

[gilmoure.livejournal.com]

So, if you throw out objective data, um, how are you supposed to judge CEO performance? I mean, is a CEO now supposed to be an artiste? Is business really not about the numbers but about more ethereal concerns? Maybe it's about the waste paper baskets?

[wcg.livejournal.com]

I think you're right, but then I'm an experimentalist rather than a theoretician.

I don't know of any sound economic reason to have the highest paid CEO. However, I can see why business executives would want people in the business culture to believe that there's a sound reason for it. Otherwise CEO jobs would be outsourced to India and Singapore just like other jobs that require good education and concentrated dedication to one's work.

[jehanna.livejournal.com]

Sounded like somebody was feeling threatened on some emotional level about the discussion relative to their feelings about money and politics, when you were trying to have a discussion of empirical evidence.

Alas, policies and convictions are so much more often the result of feelings than of unbiased analysis.

[fauxklore]

My understanding of the logic is that one needs to pay a CEO highly so that other companies won't poach him or her There are a limited number of people who are (rightly or wrongly) perceived as qualified to be CEOs of major corporations, so really this is supply and demand at work.

I'm uncomfortable with the idea of Congress capping CEO salaries normally, but I can see prohibiting money used for bailouts into being used for salaries over a certain level.

It is also a bit tough to figure out what they are counting, since a lot of compensation at those levels is in the form of stock options, bonuses, etc. - which are tied well to performance.

[sethg]

James Surowiecki: “It’s tempting to wonder if the sheer prevalence of enormous C.E.O. compensation packages means that they have some beneficial effect. But academics have found little evidence that higher executive pay leads to better company performance, and the recent study of three thousand companies actually found that the firms whose directors were the most well connected—and which paid their C.E.O.s most lavishly—in fact underperformed the market.”

But of course your ex-friend has determined The Truth About Economics and therefore does not need to be distracted by mere data. I've encountered this kind of glibertarian before.

[dillonpuff.livejournal.com]

I always thought it made sense when a CEO was getting the bulk of his compensation via stock. The position as a whole seems so ... pointless. I wonder what the day-to-day for most CEOs comes down to.

[metahacker.livejournal.com]

This idea has always bothered me. (Why do we pay them so much? Because that's what the market demands! But why...etc.)

(In your previous post you left out talking about the enablers -- the people who keep PAYING these salaries. So it's great you addressed it here!)

At a previous company we hired an overpaid MBA-type to be the CEO, and he buried the company. (It was doomed, but he ensured the doom.) I really didn't understand what he did, but he had that CEO quality, where people pay him big bucks to make stupid decisions that he doesn't have the evidence for. And I *know* his salary was truly ridiculous given the (lack of) health of the company. It's really embedded in the mentality of corporate America.

Many economists hate data that doesn't agree with their models, and ignore it as spurious. robertdfeinman has been in an uphill battle on this front for the past year since the topic started interesting him. (You can read some of his fevered rantings here or, with comments, here.)

CostCo is an exception to many rules, because its CEO demands that it be so. Natural market pressures push toward ridiculously-compensated CEOs, I think. But since it's such an exception, model-based economics ignores it as an outlier, rather than studying it as a goal. I think that might be what your flouncer was reacting to. Lacking any actual formal science, and faced with evidence that disrupts your fragile model, it's hard to take criticism. I guess.

[cheshyre]

Mind you, given the stock market's preference for short-term performance over long-term stability, that's not necessarily the best measure either.

[gilmoure.livejournal.com]

It was cool that Steve Jobs only took a dollar a year while he brought Apple back. 'Course, now, they've had some stock option scandal so I guess things do change.

[metahacker.livejournal.com]

A bit of a stunt, given how much he was earning at the time in options and other benefits, but yeah.

[devvieish.livejournal.com]

Perhaps it's circular logic? i.e. "We pay them so much, they must be worth this much." Or even "Other people have paid them, or similar CEO's so much, so they must be worth that much"? Because copying others is a common thing to do when you have absolutely no clue as to how to behave otherwise.

This is a kinder hypothesis than my first cynical thought, which is: "It's the same reason managers always get paid more - it's cause they set the salary." Or in the case of CEO's, they personally might not, but they are friends or at least in the same social networks with people like the board members who *do* set salaries.

issues in CEO pay

[(Anonymous)]

Executive pay is a tough economic problem, and nobody has an entirely satisfactory solution. The basic issue is the principal-agent problem: how should the compensation be structured to align the CEO's interests with the shareholders'? If the CEO owns the entire company, then there is perfect agreement, but of course that defeats the purpose of selling stock. The next best thing is to give the CEO a lot of stock options, so that he/she has a powerful incentive to raise the stock price.

However, that solution is economically imperfect (quite aside from any issues about fairness). The shareholders are investing in the company as part of a diversified portfolio. They would like the companies they invest in to take sensible risks, with the understanding that it might not work out in any given case, but on average they will profit. However, the CEO may be much more risk-averse, since he/she lacks the benefit of diversification. If the CEO takes a sensible risk that doesn't in fact work out, then not only are the stock options worthless, but the CEO's job and future employment opportunities are in jeopardy (since it's hard to prove in hindsight that the risk was sensible). So just giving stock options makes the CEO far more risk-averse than the shareholders would like.

This leads companies to offer generous payouts for unsuccessful CEOs. The idea is that the CEO will feel confident in taking dramatic action. If it works, the CEO will become very rich, while if it fails he/she will just be moderately rich.

Ultimately, I believe this is a pretty sensible idea. It's far from perfect, but it's not clear how to reliably improve it (cases like Sinegal at Costco don't seem easily generalizable). The biggest weakness is that it's easy to screw up: nobody really knows how to set CEO compensation, so it's mostly left up to the discretion of the board. Some boards just aren't good at it and get the incentives wrong by accident. In other cases, the boards aren't even trying to get it right, but rather to funnel money to their buddies.

There are also various psychological effects. For example, one function of enormous compensation packages is to motivate lower-level executives. For every overpaid CEO, there are 50 ambitious junior executives who work like hell to get a shot at someday becoming CEO. Some of that work is wasted (for example, in political scheming), but some of it is genuinely valuable for the company. Even if the winner doesn't deserve the reward, the overall motivational effect may be worthwhile for the shareholders.

Another psychological effect is that boards find it very difficult to reduce CEO compensation. When the company is doing well, the CEO claims credit and asks for more money. When the company is doing poorly, the CEO is usually replaced and the new CEO wants at least as much money as the previous one got. (After all, the board is bringing in the new CEO because they think he/she will be better than the previous one.) This is illogical, but it's human nature.

Anyway, all this deals with why CEO compensation is high, but only the last paragraph addresses why it has increased so much. One more fundamental reason is that the biggest companies are getting bigger, and the economic analysis suggests that executive compensation should therefore increase. The more money is at stake, the more the principal-agent problem matters and the more the shareholders should be willing to pay to ameliorate it.

One could also argue that it's the past that was exceptional, not the present. The modern economic system was just getting off the ground early in the 20th century. The great depression and WWII were severe interruptions, and there's no reason to think the 50's and 60's were really typical. From this perspective, perhaps CEOs were always destined to make a lot of money eventually, and the real question is why it took so long in coming. (Perhaps it accelerated when board members with pre-WWII business experience started to become minorities on boards.) I really don't know, but things sound very different depending on how one frames the question.

[holyhippie]

How I understand it, there are three general ways a employee can be compensated - salaries, cash bonuses, and various forms of stock compensation.

The problem with horribly overcompensated CEO's comes in the first two forms. Those sap cash out of a business, and look very bad when a that is very out of proportion to what the rank and file employees get. (Personal example: the executives of Sun Microsystems have been taking huge salaries an bonuses, while the market value of the company has gotten frittered down to be just about zero - the stock is worth just about the value of the cash on the balance sheet. They RIFed me two years ago.)

Meanwhile, it's hard to argue against CEO compensation that is exclusively tied to the value of the stock. Where I work now (Apple), the CEO has taken a salary of $1 a year for about the past ten years. He's made lots and lots of money while working there - but as far as I can tell, it's all been from various forms of stock grants.

[holzman.livejournal.com]

Your friend is full of shit. There are plenty of ways to measure CEO performance.

Stock appreciation by quarter or year or N years.
Average stock volatility over relevant period of time.
Average share dividend.
Revenue. Profit. Staff retention. Market share.

Which one to use? That depends on which one the shareholders value, which should drive what to look for in a CEO. If you can't quantify it, you can't measure it.

The real reason CEO compensations are so high is that CEOs tend to be on boards of directors for a variety of companies, therefore friends with one another, therefore motivated to give one another lots of money with poorly defined accountability criteria in order to stave off shareholder law suits.

They said that it's hard to quantify CEO performance, because someone could do really brilliantly, but, for circumstances outside their control, still lose billions in a quarter.

This is bullshit on so many levels. If the company lost money, the CEO didn't do brilliantly. Period. Even if the circumstance was outside their control, the fact that they lost billions as a result means they botched the risk management. The fact that gullible idiots will make excuses for them means they're less incented to do the risk management right.

Make no mistake that the rank and file don't get this level of sympathy -- if the company loses money, bonuses don't happen and the rubric is "We're a team, as a team we didn't earn the bonus."

The fundamental principle your ersatz friend is discussing -- which he will never cop to -- is that an unspoken assumption in capitalism is that the guys with the capital are inherently superior to the guys with the labor and are deserving of a better deal for no reason other than they're the guys with the capital.

[caprinus.livejournal.com]

. . . the weird thing is that, to me, actually looking at real-world data is kind of PART of discussing fundamental principles. I thought that starting from dogma, working outward, and deciding that THAT was reality, without actually CONSIDERING reality went out of fashion when people noticed that Aristotle had been wrong about the number of teeth that women had.

And this is why you're not an economist! Thank god.

The person said that it was to attract the best CEOs.

I said that there's no correlation between CEO compensation and CEO performance.

They said that it's hard to quantify CEO performance, because someone could do really brilliantly, but, for circumstances outside their control, still lose billions in a quarter.

I said, okay, there's no demonstrable correlation.

They said, c'mon, not all things that are demonstrable are quantifiable.

I just happened to be looking at this article... http://jama.ama-assn.org/cgi/content/full/299/9/1016

Basically, people feel something is better when it costs more. I've seen it anecdotally shown before, but this was a nice double blind confirmation. (It won an IgNoble in 2008, that's why I found it). I wouldn't be surprised if the external perception of the performance of a CEO was similarly, in the absence of easily quantifiable benchmarks, dependent on a costly placebo effect.

[ailbhe]

I want to be a CEO for a year, get one year's salary, and live off it for a LONG TIME while I do exciting and worthwhile volunteer work somewhere else. I will never understand why that's not what happens with stupidly large sums of money.

This is why I'm not a CEO.

[sethg]

I would qualify this slightly--if the company lost money relative to other companies in the same or comparable industries, the CEO didn't do brilliantly. If the company did well in a year that just about every stock on the exchange was going up, the CEO can't claim to possess any special genius, either.

[dancing-kiralee.livejournal.com]

. . . the weird thing is that, to me, actually looking at real-world data is kind of PART of discussing fundamental principles. I thought that starting from dogma, working outward, and deciding that THAT was reality, without actually CONSIDERING reality went out of fashion when people noticed that Aristotle had been wrong about the number of teeth that women had.

In a previous post somewhere, you talk about different levels of proof, starting with mathematical, or logical, proof which you consider the strongest, and moving on through scientific proof (next strongest) to various standards of legal proof - of which "beyond a reasonable doubt" is the strongest, but - and this is the point you were trying to make - there are others, and each of which has a place in the legal process.

And there is something I've always wanted to say about this... not so much about the different levels of legal proof, but about the idea that mathematical proofs are stronger than scientific ones. And this seems like a good oportunity to say it.

All mathematical proofs are based on starting from dogma (Mathematians call them axioms, because they can't be proven and have to be taken on faith), working outward using logic, and deciding that THAT is reality, without actually CONSIDERING reality.

Logic is a really good basis to use for working outward; that's why so many people are convinced that mathematical proofs are unbreakable; but ultimately it is all based on dogma/ the axiomitic method, and, since mathematical proofs are only checked for logic, and not for correspondence to the real world, well, they aren't guaranteed to match the real world - that is to be true in an imperical sense - although clearly they are true within the axiomtic system you are using to construct them.

So I've always felt that scientific methods of truth testing were stronger than mathematical ones. Science does test it hypohesises against the real world. It also uses mathematics to model the world, and while there is no way to 'prove' that model is accurate in the pure mathematical/logical sense of the the word 'prove' it is good enough to harness a great deal of the power of mathematical proof and use it as fuel for the engine of science. These two things combined make it a better means of measuring what is true in the real world, even if mathematical proofs are stronger in the abstract world of mathematics.

And, ultimately, it's the real world that counts.

If you don't mind being engaged in a (logical) debate, what do you think?

Kiralee