In theory a business that was paying somebody $8 to create $800 of value would have an incentive to pay a lot of people until either the $8 increases or the $800 decreases. If they didn't do it, somebody else would.
Vice versa a business that paid someone $8000 to get $800 worth of value might be able to sustain it (it can divert $7200 from something else) but won't be able to scale it.
People who want to sound reasonable might talk about figures like $790 and $810 but that guarantees you will be in the weeds because you never know exactly where to "charge" things. For instance, a firm invests in equipment for a worker, worker B fills in the gap when worker A slacks off, etc. Even in areas where it is widely believed you can measure productivity (e.g. sales) this is problematic.
I have been looking in the macroeconomics literature but haven't yet found any exploration of the following:
what if an organization has a nodal structure such that at each node somebody can divert X% (say 1 for the hell of it) of what goes through the node? The salary of the top man is then basically proportional to the size of the node, so to get mega-rich, a bank president in New York might buy a bank in New Jersey. The node in New Jersey remains, but a buyer of the bank in New York now gets to skim additional cream.
A corollary of that could be that the "competitive advantage" of "productive" urban areas is in theft. In Ancien Regime France you would set up at Versailles because that is where you had connections for thieving. In post-2008 America, Washington D.C. has boomed because it has "comparative advantage" if you want to bribe politicians to help you steal. Certainly in a country like Equatorial Guinea you are going to settle in the capital because that's where you can steal.