From Pattern Recognition to Economic Disequilibrium

Abstract

This article is focused on the economic works of the Soviet machinelearning pioneer Emmanuil Braverman, who published, during the 1970s, a series of papers introducing disequilibrium fixed-price models of the Soviet economy. This highly original theory, developed independently from the Western analyses of disequilibria, proposed rationing mechanisms capable, under some conditions, of bringing a system to the state of equilibrium. However, in a fixed-price economy, equilibria are not necessarily optimal or effective; therefore specific observational and analytic procedures aiming at bringing a system to a better state had to be invented. Braverman interpreted this analytic framework as a “qualitative system of control” of the Soviet economy representing a sort of a third-way solution between neoclassical models of spontaneous coordination of autonomous agents and theories of optimal planning. This innovative approach, very different from the styles of reasoning in mathematical economics of his time, was grounded in his work on pattern recognition and informed by a cybernetic vision of control as information processing and communication in complex systems.