The Life-cycle of Concentrated Industries

Are competition policies designed for static industries suitable for innovative industries where dynamic competition for the market is key? If not, how should policies differ? We develop a model of the life-cycle of an oligopolistic industry: a version of Jovanovic and MacDonald (1994) with a finite number of firms. The equilibrium features a period of intense entry, followed by a shakeout and eventual industry concentration as some firms scale through innovation while most exit. We analyze the second-best problem of a government subsidizing small firms to promote competition. Innovation and dynamic competition do not necessarily justify intervention, as the equilibrium can still be second best. In general, the optimal policy depends on the nature of competition. Firms primarily compete for the market when innovation leads to large differences in scale. The government can wait to intervene in this case; committing to do whatever it takes to promote competition if and when the industry concentrates excessively. Subsidies early in the life-cycle are unnecessary. These results contrast with calls for aggressive ex-ante regulation in highly innovative industries, suggesting a wait-and-see approach may be preferable. We apply these insights to digital and AI industries in the U.S. using data on venture-backed firms.