Phil Rosenzweig likes criticizing other people. He is usually correct but at his worst he is highly selective in his criticisms. He seems starstruck by CEOs. He happily agrees with dubious advice from top managers but gets out his microscope when examining academic writing.
In Left Brain, Right Stuff he dissects academic overconfidence research for, amongst other things, a failure to define overconfidence clearly. Many points are well made but Rosenzweig isn’t half as careful elsewhere.
What bias means to him isn’t clear. In the book he supports the, frankly extremely silly, “bias for action” so popular with macho strategy writers. In evidence for this he tells us the story of a team in a stock picking competition. Their “bias for action” meant they took a “risky” strategy because they realized that the prize was only for coming first. In such a situation, however, any expected value model would recommend such a “risky” strategy; better to risk all and maybe you win than play sensibly and guarentee second place in a game where only first place gets a prize. Thus taking an “aggressive and sometimes extremely risky strategy” (Rosenzweig, 2015, page 75) surely can’t be evidence of bias. It is just a reasonable response to the incentives and would be adopted by any “unbiased” decision maker. Rosenzweig even implies this in saying that the team showed “an astute understanding of the competitive context”(page 75).
If bias is another word for a clear understanding of the incentives in this market it is hard to reconcile with the idea that “cognitive biases are unconscious”(page 75). Rosenzweig realizes this and so he just abandons any attempt to properly define bias. He rips into overconfidence bias research for not being specific enough about what overconfidence means but won’t specify what bias itself means.
Rosenzweig defends a bias for action by illustrating how unbiased behaviour is good. Not his best moment.
Read: Phil Rosenzweig (2014), Left Brain, Right Stuff: How Leaders Make Winning Decisions, Public Affairs.