Narrow Bracketing and Dominated Choices

IZA Logo

IZA Seminar

Place: Schaumburg-Lippe-Str. 9, 53113 Bonn

Date: 26.06.2006, 12:15 - 13:30


Presentation by 

Georg Weizsäcker (DIW Berlin)


An experiment by Tversky and Kahneman (1981) illustrates that people's tendency to evaluate risky decisions separately can lead them to choose a portfolio of assets that is first-order stochastically dominated by another available portfolio. We investigate the generality of this effect, both theoretically and experimentally, and argue that Tversky and Kahneman's example is not contrived: Any decision maker who narrowly evaluates her decision problems one by one is subject to making dominated choices. Even with only a weak tendency to make choices in isolation - as opposed to taking a fully broad view - it is straightforward to construct examples where the decision maker chooses dominated portfolios. We also provide a result on the amount of money that the agent can lose, due to one single mistake of this kind. The theory is accompanied by two experiments, one for monetary stakes in a standard laboratory setting, and one with hypothetical questions and with a large sample from the general U.S. population. They provide further evidence on dominated choices, and demonstrate that a manipulation of the presentation can reduce the bias. The data also allow to describe the potential size of losses, for the average decision maker.

Download complete paper   
For more information, please contact