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 proce

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.