Distributed expertise and instructions

Abstract

Purpose – The idea of Simon according to which instructions are premises for further decisions is formalized with respect to informationally decentralized organizations. The paper aims to discuss these issues. Design/methodology/approach – The paper is built on the economic theory of teams by Marschak and Radner. Team theory is particularly suited to the analysis of informationally decentralized systems. Teams are composed of agents who strive for a common aim, but are informed of different parameters that characterize the state of nature. In particular, the members want to maximize the team payoff function that depends on both the actions of the members and the state of nature. Findings – For teams à la Marshack-Radner, since every member needs to compute optimal action rules, the paper shows that two types of expertise are required: forecasting expertise (as competence over the randomness of the parameters that affect the team payoff function) and organizational expertise (as comprehension of the relevant characteristics of the other members). When the members are supposed to have complete expertise, the only role of orders is the communication of information about the state of nature. Instead, when instructions report actions under the control of the senders, the team population can be classified into ranks, each one having a particular computational problem to solve. The paper analyzes the conditions under which instructions can correspond to different and ordered degree of intelligibility of the team operations. Through instructions non-decodable information is built directly into the decision rules chosen by the receivers, and instructions can realize economies in the use of expertise. Originality/value – Instructions are formally shown to play a crucial role when expertise is differently distributed across agents.