Aggregation in Decision Problems: Concepts and Applications

This paper discusses the concept of aggregation in decision problems with the Bayesian approach. A variety of examples are provided for illustrative purposes. An aggregation error is said to occur when analyses made at aggregate and disaggregate levels yield different re suits. In the absence of aggregation error, perfect aggregation is said to occur. Perfect aggregation is shown to be almost impossible and consequently aggregation error is practically inevitable. Alternative measures of aggregation error are provided. Also, the impact of aggregation error on the decision to be made is analyzed.