AbstractThere are many computational models available now which purport to add to our ability to plan and control organizations in a decentralized fashion. Some of these are custom built for the purpose while others incorporate mathematical programming decomposition procedures. Although comparisons have been made concerning their computational requirements, little analysis seems to have been done to explore the organizational implications of the various methods. A start is made on this task here and some conclusions drawn in terms of the managerial style of-and the motives for-decentralization.