A two-stage model of research and development with endogenous second-mover advantages

This paper describes a simple two-stage model of research and development, in which the ‘winner’ of the research stage has the option of moving first in the development stage. Some unexpected results emerge: in equilibrium, the leader in the development stage invests less than each follower, and is consequently least likely to collect the patent. Moreover, the leader receives a lower expected payoff than each of the followers. Thus there are endogenous second-mover advantages. Using a game of timing (in which the identity of the Stackelberg leader is determined) to link the two stages, I find that firms face quite different incentives in the research stage. Although the leader invests less than each follower in the research stage as well, the leader enjoys higher expected revenue from the complete (two-stage) game than does each follower. The equilibrium is inefficient because there is a lag between the time at which research is completed and the time at which development is begun, and because aggregate investment is inefficiently (asymmetrically) distributed across firms.