This article develops a method for drawing samples from a distribution with no finite quantiles or moments. The method provides researchers with a way to give subjects the experience of ambiguity. In any experiment, learning the distribution from experience is impossible for the subjects, essentially because it is impossible for the experimenter. We characterize our method, illustrate it in simulations, and then test it in a laboratory experiment. Our method does not withhold sampling information, does not assume that the subject is incapable of making statistical inferences, is replicable across experiments, and requires no special apparatus. We compare our method to the techniques used in related experiments that attempt to produce an ambiguous experience for the subjects. This paper was accepted by Peter Wakker, decision analysis.ambiguity, Ellsberg, Knightian uncertainty, laboratory experiments, decision analysis, theory
We study a disclosure decision for a firm's manager with many sources of private information. The presence of multiple numerical signals provides the manager with an opportunity to hide information via aggregation, presenting net amounts in order to show information in its best light. We show that this ability to aggregate fundamentally changes the nature of voluntary disclosure, due to the market's inability to verify that a report is free of strategic aggregation. We find that, in equilibrium, the manager fully discloses if and only if the manager's private information makes the firm look sufficiently weak. By separating bad news from good news, a disaggregate report informs the market of as much offsetting news as possible, revealing how close the news is to a neutral benchmark. The result is, therefore, pooling at the top and separation at the bottom, the opposite of what transpires with a single news source.
JEL Classifications: M41; D82; D83.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.