The Ratchet Only Turns Up

In 1965 the boss of a big American company made about twenty times what one of his workers made. Last year he made two hundred and ninety times as much.

Not twenty percent more. Two hundred and ninety times. Over the same stretch, the pay of the person on the floor rose 24% after inflation, while the pay at the top rose more than a thousand. Same economy, same country, same tax code roughly — two completely different rides.

There are two stories about why, and each has a powerful lobby.

The first is the one you hear in boardrooms, and it isn’t stupid. Call it the superstar story. A modern corporation is enormous, and the person steering it sits on a lever with a very long arm. If a chief executive is even one percent better at allocating capital at a company worth a hundred billion dollars, that one percent is a billion dollars. Against a number like that, a twenty-million-dollar pay package isn’t a scandal — it’s a rounding error, and a bargain if the person is any good. The economist Sherwin Rosen wrote the math of this down back in 1981: in a big enough arena, tiny differences in skill translate into gigantic differences in value, and therefore in price.

Take that seriously. It’s the strongest thing the defenders have, and it’s genuinely true as far as it goes.

Here’s how far it goes.

If pay were really the market pricing scarce talent, a few things should follow, and you can check them. Higher-paid CEOs should run better companies. They mostly don’t — once you adjust for the size of the firm, the link between how much a chief executive is paid and how well the company then does is weak to invisible. The most exhaustive account of this, Lucian Bebchuk and Jesse Fried’s Pay Without Performance, landed in 2004 and has not really been answered: boards, they showed, don’t bargain with executives at arm’s length. They bargain across a table the executive helped build.

And then there’s the single cleanest tell in the whole argument. The ratchet.

Here’s how executive pay actually gets set. The board hires a compensation consultant. The consultant picks a group of “peer” companies and checks where the CEO’s pay sits against them. And essentially every board, everywhere, wants its CEO paid at or above the median of that peer group — because who wants to admit their leader is below average?

Sit with that for a second.

Everyone paying above the median. It’s the town where all the children are above average. It cannot be true for all firms at once — the median is, by definition, the middle — so the only way everyone can be “above” it is for the median itself to climb, every year, forever. That’s not a market finding a price. A market clears; it settles. This escalates. A market that only ever moves one direction isn’t discovering value. It’s a ratchet, and a ratchet has a name for the little catch that stops it slipping back: a pawl. Executive pay grew a pawl sometime around 1985 and has been clicking upward ever since.

Now the part that should end the “it’s just the market” story for good.

Private equity firms buy companies and run them with real owners breathing down the CEO’s neck — a few sophisticated people with their own money on the line, watching everything. If the superstar story were the whole truth, they’d pay the going rate for that scarce talent, same as anyone. They don’t. They pay their CEOs noticeably less — often a third to a half less — for running companies of comparable size. Same talent pool. Same labor market. The only thing that changed is who’s watching the pay decision. When the owners actually watch, the “market price” quietly drops by half.

So is it all a con, the superstar idea pure ideology?

No — and this is where the people who hate CEO pay overreach. The superstar mechanism is real. Firms are vastly bigger and more tangled than they were in 1965, and the person at the top does sit on a longer lever than his grandfather did. That justifies a chief executive being paid more than twenty times a worker. It sets a floor. What it cannot do is explain the level — why the number is 290 and not 60, and why it’s 290 in New York and a small fraction of that in Tokyo, for running companies that compete in the very same global markets. If a single world market for talent set the price, the prices would converge. They don’t even come close.

Which means the honest answer isn’t “market” or “governance.” It’s both, doing different jobs. The market explains why CEO pay went up. Governance failure — captured boards, the ratchet, a tax code that in 1993 shoved everyone toward stock options, oversight votes so toothless that fewer than three in a hundred ever fail — explains why it went up so much.

But there’s a deeper question hiding underneath, and it’s the one both camps flinch from. The principal-agent problem — bosses gaming owners who can’t watch them closely — didn’t appear in 1985. It was every bit as available in 1955, when the ratio sat quietly at twenty to one. Weak governance was always possible. So what changed?

A norm broke. For a generation after the war, there was a real, enforced sense — in the press, in politics, inside boardrooms themselves — that a certain gap was simply unseemly. Peter Drucker, no socialist, floated twenty-to-one as the point past which a company started to eat its own cohesion. That restraint wasn’t a law. It was a manners, and manners are fragile. Somewhere in the 1980s it dissolved, and once it was gone the ratchet had nothing left to push against. That’s the uncomfortable finding for economists on both sides: part of what sets the pay of the most powerful people in the economy isn’t leverage or agency costs at all. It’s whatever the room currently considers shameful.

The inequality angle is the part everyone reaches for, and it’s true but subtler than the slogans. CEO pay is not, by itself, the biggest engine of the top one percent’s rise — capital and business income do more of that lifting. What CEO pay is, is the visible engine, and a contagious one. When the chief’s number climbs, the pay of the next dozen executives climbs behind it, benchmarked to the same rising median, and the whole top of the building floats up while the floor stays put. The ratio isn’t just a statistic about one person. It’s the shape of the modern firm.

So, next time a board tells you its CEO is simply worth every penny of a market price: ask them why the same person would cost half as much the moment someone with real skin in the game was actually watching.

And ask what number, exactly, would finally strike them as too much — and who, in that room, is left to say it.