Organizations and Markets

Overconfidence

26 July 2010 at 12:56 pm Peter G. Klein 5 comments

| Peter Klein |

Busenitz and Barney (1997) famously argued that entrepreneurs (founders) are particularly susceptible to overconfidence and representativeness biases. Compared to professional managers, entrepreneurs systematically overestimate the probability that a new venture will succeed and tend to draw unwarranted generalizations about the future from small samples. Overconfidence is now one of the major themes in the contemporary entrepreneurship literature (Bernardo and Welch, 2001; Forbes, 2005; Koellinger, Minniti, and Schade, 2007).

A new NBER paper by Itzhak Ben-David, John Graham, and Campbell Harvey finds evidence for a particular kind of overconfidence, “miscalibration,” among corporate executives. Miscalibration occurs when the agent’s forecast probability distribution is too narrow, meaning that the likelihood of extremely positive or negative events is unrealistically discounted. The idea is that agents with miscalibrated expectations are overconfident, not in the success of their activities (what the authors label “optimism”), but in their ability to predict the success of their activities. Survey evidence from a sample of CFOs reveals a number of interesting regularities about the relationship between miscalibration and past financial performance, corporate investment, and other observables. Here’s the abstract:

Miscalibration is a form of overconfidence examined in both psychology and economics. Although it is often analyzed in lab experiments, there is scant evidence about the effects of miscalibration in practice. We test whether top corporate executives are miscalibrated, and study the determinants of their miscalibration. We study a unique panel of over 11,600 probability distributions provided by top financial executives and spanning nearly a decade of stock market expectations. Our results show that financial executives are severely miscalibrated: realized market returns are within the executives’ 80% confidence intervals only 33% of the time. We show that miscalibration improves following poor market performance periods because forecasters extrapolate past returns when forming their lower forecast bound (“worst case scenario”), while they do not update the upper bound (“best case scenario”) as much. Finally, we link stock market miscalibration to miscalibration about own-firm project forecasts and increased corporate investment.

I’m not aware of any entrepreneurship studies that distinguish miscalibration from optimism, in the sense those terms are used here. Am I missing something?

Add to: Facebook | Digg | Del.icio.us | Stumbleupon | Reddit | Blinklist | Twitter | Technorati>

Entry filed under: - Klein -, Entrepreneurship, Financial Markets, Management Theory, Strategic Management.