Technological Change and Agency Costs. A Growing Dilemma for Soviet Central Planners. Growth rates in the Soviet economy have slowed significantly in recent years. It appears that input growth rates cannot be increased substantially, so the burden of reversing the retardation falls, in large measure, on technological change. Yet the Soviet economic system is notoriously weak in generating and diffusing improved technology. These weaknesses are well known and yet they persist in the system. The central argument of this paper is that these systematic weaknesses are not necessarily a reflection of irrationality or ignorance, but rather reflect the results of an implicit trade-off between the objectives of economic efficiency and control of the system. Taking off from Zaleski's conclusion that the purpose of the planning system is to manage, the papers's analysis applies recent developments in the theory of agency costs to show why apparently illogical or irrational characteristics persist in the system. For example, the policy of permanent prices, the proclivity for long production runs of standardized products, and the reluctance to make allowances for obsolescence, among other things, can be understood in these terms. The basic analysis is then applied to characteristics of the system that appear to slow technological change. A main conclusion is that the existing planning system can be better understood as an instrument that is used to promote the ends of the leadership. To be effective, the system must be controlled, and it is the effort to maintain control that results in some of its apparently anomalous characteristics. In trying to understand the system or reform attempt, it is important to evaluate how well the system serves the ends of the leaders, rather than whether it conforms to neoclassical economic welfare criteria.