Not sure how relevant all that is.
A cap would result in an immediate Posh-Boy Drain. The question is whether we'd notice any difference to performance. At the moment shareholders don't really have any stats on how important high pay is.
What does company X offer to executive A to get him/her to leave company Y if executive A is already at max salary? Fewer hours? More hidden perqs?
The key element of the statistical distribution of salary isn't the level of the maxima, but the existence of differentials. Given any two people with different skills, experiences, accomplishments, personalities, etc. what is the difference in their salaries? Under what conditions should person A earn just 10% more than person B? What's the reasonable differential for a year of experience, a college degree, people skills, a college degree from a good school, or a PhD? And what if a person shows quantitative differences in productivity relative to their peers? If one extraordinary person can produce as much as two ordinary people, wouldn't the extraordinary person expect more pay?
It doesn't take many reasonable 10% pairwise differences across a working population of millions or billions of people to create a 1,000,000:1 ratio between the least-paid and most-paid person. It's the chaining of these reasonable differentials that drives the extrema.