Salmon points out that “not a single one of the 12 [candidates] is a CEO who was hired to run a company by its board of directors.”
Let me give an alternative explanation for this. Founder CEO's take more risk. And we would expect risk takers to be overrepresented in outlier categories such as "best ceo candidate".
The same could potentially explain "analysis of recent exits for high technology companies". They found that founding ceos make better exits. Well that has a survivorship bias, because companies that didn't make an exit (because of a gamble gone wrong?) are not included.
EDIT: Drum's post suffer from the same problem. If you evaluate performance during [t0, t1], to not get survivorship bias, you need to have an inclusion criteria based on only information known at t0.
Let me give an alternative explanation for this. Founder CEO's take more risk. And we would expect risk takers to be overrepresented in outlier categories such as "best ceo candidate".
The same could potentially explain "analysis of recent exits for high technology companies". They found that founding ceos make better exits. Well that has a survivorship bias, because companies that didn't make an exit (because of a gamble gone wrong?) are not included.
EDIT: Drum's post suffer from the same problem. If you evaluate performance during [t0, t1], to not get survivorship bias, you need to have an inclusion criteria based on only information known at t0.