> By analyzing the past returns of current S&P500 companies, the author is allowing for survivorship bias;

Thought so myself. Interesting to see what the result were with data from the beginning of the period.