Modelling Agent Forecast Distributions

Mike West
Duke University

Aug 31 1989

Problems in the analysis and interpretation of uncertain inferences obtained from individuals, or agents, are considered in the context of forecasting a scalar random quantity. A new theoretical framework for such problems is described here. Bayesian updating is developed for cases in which agents' opinions are provided in terms of full forecast distributions, forecast quantiles and other partial specifications. Example models highlight the key features of the theory, and illustrate application.


agent opinion, combination of forecasts, expertise


PDF icon 1989-12.pdf