I'd say it is basic game theory and statistics.
If we have population of individuals who exhibit three types of behaviour:
A - altruistic (1% of population)
E - egoistic (1% ...)
N - Normal (randomly A or B 98% of population)
Let's them play a game where E - type has muuuuch better chances to win....
It is obvious that the proportion of E - types in the winner population will grow during the tournament.
This explain the current state of world affaires: Usually 'Supreme Leader' is the "biggest bastard of them all"...
It's basic extrapolating-without-data is what it is. You have a theory, well, let's call it a hypothesis ("a person who reaches CEO level is someone who will not behave altruistically"). You even have a simple model which mostly agrees with the hypothesis (not completely, I'll note) and which relies on several more untested hypotheses. Both the hypothesis and the model predict certain things about the world. You can check! Are those predictions borne out in reality? (˙˙˙ʎʃʃɐǝɹ ʇoᴎ :ɹǝʍsu∀)
