ESSAYS ON INNOVATION, PRICE DISPERSION, AND MARKET POWER

Abstract

This dissertation covers three chapters on innovation, price dispersion, and market power. In Chapter 1, I show that technologies that are equally green can yield very different gains in social welfare. I examine whether market-driven innovation in green industries follows the socially optimal technological path when firms face competing options with distinct cost structures, and explore policies that can correct potential inefficiencies in path choices. I estimate a dynamic structural model of Chinese electric vehicle (EV) battery suppliers choosing how much to innovate along two competing paths: Lithium-Ion–Ferro–Phosphate (LFP) and Nickel–Cobalt–Manganese (NCM). The analysis shows that a social planner would undertake about four times as much innovation as the market in LFP, which involves higher innovation sunk costs but delivers lower marginal costs in production. By contrast, the market innovates about twice as much as the social planner in NCM, which has the opposite cost structure. I attribute this divergence primarily to vertical separation between battery suppliers and EV makers, competition between EVs, and limited demand for EVs in early years, rather than to innovation spillovers or environmental benefits that firms fail to internalize. Finally, I demonstrate how R&D subsidies could help correct the path choice inefficiency and shift innovation toward the socially preferred path. In Chapter 2, my coauthor, Yihui Fu, and I study how airline competition reshapes price dispersion and which travelers benefit most from entry. Using novel booking-level data from China’s domestic airline industry that record the booking date, departure date, and departure time of every ticket sold, we decompose aggregate price dispersion into three components: across the booking horizon, across departure times of day, and across departure dates. Our estimates reveal sharply heterogeneous effects: competition compresses dispersion across departure times, leaves booking-time dispersion essentially unchanged, and marginally widens departure-date dispersion. Because the departure-date dimension dominates the aggregate, the Borenstein and Rose [1994] and Gerardi and Shapiro [2009] frameworks yield the same positive aggregate sign in our data. Evening flyers, flexible-date travelers, and late bookers capture the largest gains from entry. High-speed rail availability amplifies the widening on short routes, consistent with the industry elasticity effect predicted by Holmes [1989] and Stole [2007]. In Chapter 3, my coauthor, Bruno Pellegrino, and I examine the joint welfare effects of monopoly and monopsony power in the U.S. economy, where firms often exert influence in both product and factor markets. We develop a general equilibrium framework that captures the network of firm interactions through supply chains, product competition, and labor markets. Using data on publicly traded firms, we estimate that combined market power reduced total surplus by 5.7% in 2015. Importantly, monopoly and monopsony effects are sub-additive. Furthermore, we find that accounting for monopsony in factor markets not only exacerbates the welfare damages of hypothetical mergers but also renders Cournot competition welfare-superior to Bertrand, highlighting the nuanced interplay of dual market power.