To Gangster Octopus’s "why’s that"s, the answer is simple. There is no reason to believe that this relationship exists. We’ve shown many reasons why many CEO’s are compensated so richly. All of them point to an inefficient marketplace, where it is in the best interest of everyone making CEO pay decisions to increase this pay.
To the second point, when you state that poor performance can occur “for many reasons that are not the CEO’s fault”, I agree. In the same manner, good performance also can have very little to do with the CEO. So, why pay them so much money?
As for cites, there are lots. I can’t find any studies that I’ve analyzed in this regard on the web, but Google will lead you hundreds. And the lion’s share all show that there is no relationship between CEO pay and shareholder return.
http://links.jstor.org/sici?sici=0019-7939(199002)43%3A3<13S%3AEPAFP>2.0.CO%3B2-G&size=LARGE#abstract
http://www.baselinemag.com/article2/0,1540,1990245,00.asp
In reality, there is tons of research available, if you look for it.
To Cheesesteak, pointing out that several CEO’s have had success doesn’t answer the question. Just based on a normal distribution of performance, you’ll always find a few CEO’s who have had multiple successes, those who are 2-3 standard deviations up on the high end of the curve. This is pure luck, or rather, having the good fortune to work for a good company that is in the right markets at the right time. The point is, would the companies have done worse with a lower paid CEO? Probably not.
In general, opponents of high CEO pay will always point to rational, qualititative and quantitaitve reasons why CEO pay is way too high. Proponents always have to seem to point to ideas that can’t be proven such as “the importance of experience” and “former success”. It ends up being a debate similar to the science vs. faith debate.