> If providing something has high impact, but that something is very common, then it has low worth, irrespective of how high impact it is. People would die with

> If providing something has high impact, but that something is very common, then it has low worth, irrespective of how high impact it is. People would die without oxygen and water, but neither of these things are expensive. With employees the heuristic is (very roughly) market price + VORP (value over replacement player). If an average programmer typically adds $300k of value to your company, and market price for an average programmer is $70k, and you find one who is twice as effective-you can pay them a lot more than $140k. > The biggest reason CEO compensation is so high is because they are in a pyramid peak position where they can skim off a small percentage of a lot of people's productivity and claim it as their own. This isn't a bad thing. If the CEO can boost the average employee's productivity by 1%, that can be enormously impactful. The difference between a CEO who makes a 1% impact and a CEO who makes a .5% impact translates to huge sums in large corporations. So assuming that we had the ability to distinguish between good and bad CEOs, and that good CEOs were rare, it would certainly make sense to pay good CEOs huge sums and average CEOs much less. Unfortunately as you said, no one has gotten good at evaluating CEOs yet.