Hyunjae Lee

This paper develops a Schumpeterian growth model in which endogenous industry life-cycles shape aggregate growth. Potential entrants choose whether to join an existing industry or create a new one. Industries begin as monopolies, attract followers into a patent race, and gradually mature as innovation slows. Individual industries follow non-stationary paths, yet a stationary cross-section sustains steady aggregate growth as new industries continually replace maturing ones. This cross-section delivers an explicit decomposition of growth into variety creation, frontier innovation, and catch-up innovation. Because entrants cannot fully appropriate the knowledge a new industry adds, the decentralized economy creates too few new industries relative to the constrained optimum. Calibrated to U.S. data, the model shows that the effectiveness of innovation policy depends on how the cross-section of industries responds to this entry margin.