ABSTRACT
An example seminal to interest in uncertainty (or ambiguity) aversion is Ellsberg's (1961) “two-color”problem. There is a “known urn”which contains 50 red balls and 50 black balls, and an “unknown urn”which contains a mix of red and black balls, totaling 100, about which no information is given. Ellsberg observed (as did many afterwards, more carefully) that a substantial fraction of individuals were indifferent between the colors in both urns, but preferred to bet on either color in the “known urn”rather than the corresponding color in the “unknown urn.”This violates not only expected utility (EU), but probabilistically sophisticated behavior more generally. One contemporary criticism of the displayed behavior was put forward by Raiffa (1961) who pointed out that flipping a coin to decide which color to bet on in the unknown urn should be viewed as equivalent to betting on the “known”50-50 urn. One can think of such preferences as displaying a preference for randomization.
