Incentive-Compatible Elicitation of Quantiles

Abstract: Incorporation of expert information in inference or decision settings is often important, especially in cases where data are unavailable, costly or unreliable. One approach is to elicit prior quantiles from an expert and then to fit these to a statistical distribution and proceed according to Bayes rule. Quantiles are often thought to be easier to elicit than moments. An incentive-compatible elicitation method using an external randomization is available. Such a mechanism will encourage the expert to exert the care necessary to report accurate information. A second application might be called posterior elicitation. Here an analysis has been done and the results must be reported to a decision maker. For a variety of reasons (possibly including the reward system in the corporate hierarchy) the modeler might need the right incentive system to report results accurately. Again, eliciting posterior quantiles can be done with an incentive compatible mechanism.