Mathematical programming models have been developed to represent imperfectly competitive (oligopolistic) market structures and the interdependencies of decision-making units in establishing prices and production levels. The solution of these models represents an economic equilibrium. A subgame perfect equilibrium formulation explicitly considers that each agent`s strategies depend on the current state of the system; the state depends solely on previous decisions made by the economic agents. The structure of an industry-wide model that is formulated as a subgame perfect equilibrium problem is a matrix of simultaneous mathematical programming problems, where the rows represent time periods and the columns represent agents. This paper formally defines the subgame perfect equilibrium problem that includes mathematical programs for agent decision problems, and it characterizes the feasible space in a way that is conducive to the solution of the problem. The existence of equilibrium solutions on convex subspaces of the feasible region is proved, and this set is shown to contain the subgame perfect equilibrium solutions. A procedure for computing equilibrium solutions and systematically searching the subspaces is illustrated by a numerical example.