To get a quick understanding of this paper, ask yourself the following question: when do biases (financial or otherwise) play a large role in decision making?
Your answer probably includes "in areas where there is much subjectivity" or "in gray areas". Why? Because you can not readily prove the opinion wrong.
For example, why do most drivers feel they are "above average"? Probably because there is no easily identified (and readily available) measure of driving abilities, thus everyone is more or less allowed to think what they want and hence they are all "above average". (The reverse is true in track and field where an abundance of hard-fast (pun intended) data allows instant evidence of one's relative standing.)
Kumar ingeniously takes the above logic and applies it to financial markets. In the author's own words:
"In particular, how does value ambiguity (or information uncertainty) influence the strength of investors’ behavioral biases? Are investors more likely to make mistakes in settings with greater uncertainty or do stocks that are more difficult-to-value attract a clientele that exhibits stronger behavioral biases?"This idea in and of itself is not new, but the empirical testing of this idea is and that is exactly what Kumar does! And sure enough, he finds that investors behave as behavioral finance would predict.
From the paper:
"Consistent with the theoretical predictions, I find that investor overconfidence is higher for stocks which are more difficult to value. Additionally, I find that higher market-wide (or economy-wide) uncertainty (higher mean stock volatility, higher unemploymentrate, etc.) induces greater overconfidence among individual investors."He goes and and finds that several other biases are also more prevalent (for instance disposition bias) when uncertainty is high.
People will definitely be talking about this one! I^3! (for the new readers, I^3 is Interesting, Important, and Informative)
Kumar, Alok, "When do Investors Exhibit Stronger Behavioral Biases?" (June 2005). Available at SSRN: http://ssrn.com/abstract=738884
To add my two cents to this: These findings suggest that biases are always there, but to varying degrees. It would thus be interesting to see some of the recent behavorial finance literature re-examined over different time periods to see how much the results may change. Stay tuned!