[panic]

People aren't paid based on how much value they create. They're paid so they agree to work for you, and then so they don't quit working. There is a relationship here -- you're willing to pay more to keep people working on more valuable things -- but there are tons of other factors that affect a person's pay. Previous salaries, other offers, and negotiation skills often affect pay far more than actual created value.

Think about it: if pay were based on how much value a person creates, why do different people on the same team doing the same job sometimes have wildly different salaries? Why do people often get raises after threatening to quit? How can salary negotiation ever work?

[jecjec]

On the average of millions of privately arranged employment agreements, the employee's marginal value absolutely (slightly) exceeds their salary. You have to make a strong case that the labor market does not exhibit the properties of close to perfect competition: there are many buyers, and many sellers, of labor.

Sure, there are obviously exceptions to this calculation in both directions: how else can you explain Marissa Mayer's severance package? How else can you explain DHH's wage of $0 for maintaining Ruby on Rails?

If you can make a compelling case on this topic you can probably get published in major academic journals.

[panic]

> On the average of millions of privately arranged employment agreements, the employee's marginal value absolutely (slightly) exceeds their salary.

Are you saying this based on data? The actual labor market is very different from the abstract, perfectly competitive microeconomics model. It's hard to even figure out what the marginal value of a person's work is, not to mention the lack of perfect information, the difficulty of firing, the expectations around raises and pay cuts, people's varying self-conceptions of their own worth, …