The potential of an incumbent firm to retain its dominant position in a market is a focal point of the modem theory of industrial organization. A recent literature examines this issue for markets in which challengers threaten an incumbent's position by engaging in research for a new technology. Richard Gilbert and David Newbery (1982) show that an incumbent firm with a firstmover advantage will preempt potential entrants by inventing a new technology slightly earlier than would its rivals. In a stochastic R D therefore its introduction does not alter the decision to research. When the pre-innovation market is characterized by an incumbent and potential entrants, Stephen Salant (1984) shows that licensing ex post to innovation can alter the research decision as anticipation of the returns from future licensing encourages research by the entrant. As a counterpoint to the Salant result, I find that licensing ex ante to research may discourage further research by the entrant. Indeed, this strategic incentive for licensing may be so strong as to persuade the incumbent as well as the entrant to terminate research activity. To isolate this incentive for licensing from the one discussed by Salant, a simple model is constructed in which licensing of future technologies is not desirable. Anecdotal evidence suggests that the strategic incentive does lead innovators in some markets to license their technologies. Examples of this information sharing are prevalent in markets where technological change is rapid. In the 1940's, Standard Oil of New Jersey discovered a process for synthetic rubber. It traded patent rights on this process for a synthetic oil production process discovered by Farben, a German rubber company. In both the synthetic oil and synthetic rubber markets, rival companies were *Department of Economics, Scarsborough College, University of Toronto, Toronto, Ontario, Canada, M5S lAl. I thank Jon Cohen and Ron Saunders for discussions that inspired this paper. I am also grateful to Yehuda Kotowitz, Andy McLennan, Jennifer Reinganum, Ralph Winter, Myma Wooders, two anonymous referees, and seminar participants at the University of Toronto for helpful comments on earlier drafts.

We show that a new product monopolist may benefit from (delayed) competition if consumers incur setup costs. Setup costs create a dynamic consistency problem: the monopolist cannot guarantee low future prices once customers have incurred those costs. We show that, if customers anticipate this problem, the monopolist's profits can be improved through ex ante commitment to competition in the post-adoption market, if setup costs are large. If setup costs are small, the monopolist can typically achieve the same level of profits without price commitment as with.