Unfortunately, salary information is generally kept confidential. Nature is much more forthright with its facts.

There is some information on, say, glassdoor.com, but it's anecdotal. Is there a reliable and high-res salary datasource?

:-) Yes, but unintentionally!

What I meant to convey with that first sentence is that any effort at measuring the effect is bound to be highly politicized and subjected to endless debate, so it's best to focus first on the documented actions of these CEOs, which are objective facts.

The most relevant questions right now, I think, are not about the magnitude of the effect, but about the actions themselves: who, what, when, why, and to whom?

Maybe I'm misunderstanding you, but empirical data on human behavior is notoriously more difficult to both collect and control for than data on more basic processes. I think it's glib to suggest that the effect here should be easily measurable. I also don't think the actual effect is relevant in the first case: no reasonable ethical accounting would gauge the immorality of the act in proportion to its final impact on wages.