A shareholder at the 1998 Berkshire meeting stood up and asked the question everyone asks about executive pay: how do you justify these exorbitant salaries, bonuses and perks?

Buffett’s answer did not go where the question expected.

“In my own view, the most exorbitant are not necessarily the biggest numbers. What really bothers me is when companies pay a lot of money for mediocrity, and that happens all too often.”

— Warren Buffett, Berkshire Hathaway annual meeting, 4 May 1998 (watch from 4:25)

He went further, and it is the part that gets left out of the outrage coverage:

“I do not mind paying a lot of money for performance. It’s done in athletics, it’s done in entertainment… But I am bothered by irrational pay systems. And I’m particularly bothered when average managers take really large sums.”

So the complaint is not how much. It is what for. A CEO who turns a company into something worth many billions more than anyone else would have is worth a great deal, and Berkshire has said so repeatedly about its own managers. A CEO drawing the same package for keeping a seat warm is a transfer from you to them.

This is a more useful frame than the usual one, because “is this pay too big?” has no answer, while “is this pay tied to anything?” does.

Why the system produces this outcome

Buffett also described the mechanism, and it is worth understanding because it explains why the problem is structural rather than a matter of individual greed:

“The system feeds on itself. And companies do look at other companies’ proxy statements, every CEO does. And they say, ‘Well, if Joe Smith is worth X I have to be worth more.’”

Benchmarking against peers means no board wants a below-median CEO, so every package drifts upward toward the top quartile, and the median moves every year. Nothing in that loop is connected to results.

And the reason it persists:

“The people who have their hands on the switch are the beneficiaries of the system.”

Munger, characteristically, declined to be diplomatic about the compensation consultants who advise on all this. You can find that line yourself.

Berkshire’s own arrangement made the point without an argument: directors were paid $900 a year. As Buffett noted, he and Munger had not fully thought that through, since those same directors set their pay.

Takeaway

Stop asking whether a CEO is paid too much. Ask what they had to do to get it, and how much of their own money rides on the answer.

Two things worth measuring

Buffett’s framing splits cleanly into two questions you can actually check in the filings, and they are different questions.

Skin in the game

Does this person lose real money if the shares fall? Measured as the value of stock they hold against their annual pay — plus any recent open-market buying, which is the least ambiguous signal in the filings.

Pay discipline

Is the package rational? How much is genuinely at risk versus guaranteed salary, how large it is relative to what the company actually earns, and whether the CEO-to-worker ratio is sane.

The first one has a nuance that trips people up. Founders score well because they own a large stake, which is fair — but a hired CEO holding many years of pay in stock has the same exposure to being wrong, and should score just as well. What matters is not how the stock was acquired. It is whether losing it would hurt.

Fat Pitches combines those into a CEO Alignment score, 0–100, as two equal halves: 50% skin in the game, 50% pay discipline. Below 40 reads as misaligned, 40–70 aligned, above 70 owner-minded. Founders and owner-operators cluster high; richly paid CEOs holding little stock cluster low. It is a US-filer measure, because it leans on SEC compensation disclosure — for non-US names it falls back to raw ownership percentage, which is a thinner read and is labelled as one.

Buybacks: the same argument, one level down

Capital allocation is where alignment stops being a personality assessment and becomes arithmetic — and buybacks are the cleanest test.

Munger’s objection to buybacks was never that they are bad. It is that their value depends entirely on the price paid. Retiring stock at 8× earnings creates value for the owners who stay. Doing exactly the same thing at 30× destroys it. Same action, opposite result, and the press release reads identically either way.

So the useful question is not “does this company buy back stock?” but “at what multiple has it historically chosen to?” A management that buys more when the stock is cheap and less when it is dear is telling you something real about how it thinks. One that buys hardest into strength is telling you something too.

Worth checking alongside it: buybacks net of stock-based compensation. A company repurchasing $2bn of stock while issuing $1.8bn to employees has not returned much to you; it has funded payroll and called it a return of capital.

One honest limitation here. This kind of verdict rests on well under a dozen fiscal years of data — so Fat Pitches reports it as a band (owner, momentum, neutral, insufficient) rather than a score out of 100. A precise-looking number would imply a confidence the sample cannot support.

What incentives actually reward

The last place to look is the proxy statement’s list of performance measures — the metrics the bonus is actually paid on. This is the most direct statement a company will ever make about what it wants its executives to do, and it is frequently at odds with what the investor-relations deck says.

Some patterns to notice:

  • Earnings per share as the headline target. EPS can be raised by buying back stock, so a management paid on EPS has a mechanical incentive to repurchase shares regardless of price.
  • Revenue or “adjusted EBITDA” growth with no capital-efficiency term. Growth is easy to buy if nobody is measuring what it cost.
  • Return on invested capital, or per-share intrinsic value. Harder to game, rarer, and a genuinely good sign when you find it.
  • Total shareholder return over one year. Mostly a bet on the market.

Munger’s compression of all of this is the most-quoted line in the field for a reason: show me the incentive and I will show you the outcome. The proxy statement is where the incentive is written down.

Four checks, in the filings, in about ten minutes

  • What is the CEO's stock worth, divided by their annual pay? Under 1× is a salary, not a stake.
  • Has any insider bought on the open market recently — with their own money, not a grant?
  • Over the last decade, did buybacks happen at low multiples or high ones?
  • What are the bonus metrics, and can they be hit without the business improving?

Where this sits in the process

In the app this is step three of seven, and its position is not arbitrary. It comes after business risk and quality, because a wonderful set of incentives cannot save a business that is structurally broken — and before valuation, because who is allocating the cash flow changes what that cash flow is worth to you.

It is also the step most resistant to automation. A score can tell you the CEO holds two years of pay in stock and that buybacks happened at 24× earnings. Whether this particular management team can be trusted with the next decade is a judgement, and it belongs in the conviction checklist where a human signs their name to it.

The full walkthrough is in inside the Fat Pitches platform; the reason management risk is screened before price is in the two risks.