GANDAL, NEIL SCOTT; PHD

                         UNIVERSITY OF CALIFORNIA, BERKELEY, 1989
 
                         ECONOMICS, GENERAL (0501)
 

                         This dissertation explores three settings where the incentives for innovation are affected by
                         interdependencies among strategic agents. The first essay asks to what extent joint ventures,
                         assembled by a mechanism designer, can improve on the crude incentives for research provided by
                         patent law. In a simple model, we show that, when research qualities are unobservable, there is a
                         balanced-budget mechanism that achieves the first best, and there is also a mechanism without budget
                         balance that holds firms to their reservation payoffs (which are established by expected payoffs in a
                         patent race), while implementing first best actions. Therefore, a profit-maximizing contract giver will
                         exhaust all potential for improving efficiency. We define a mechanism for the case that qualities and
                         actions are both unobservable, and discuss when the first best can be implemented. The second essay
                         examines an environment in which an upstream innovator develops a sequence of process innovations
                         and investigates the profitability of several different licensing schemes, when there are two downstream
                         firms. If the innovating lab expects to achieve only a single innovation, it is likely that both downstream
                         firms will receive licenses. When the lab expects to achieve a sequence of innovations, diffusion is likely
                         to occur by exclusive licensing. The analysis suggests that a sequence of innovations will lead to a less
                         competitive downstream industry than a single innovation. The third essay addresses the adoption of
                         technology when there are network externalities and networks are characterized by complementary
                         products produced by different firms. Within this 'hardware - software' paradigm, we explore how the
                         pricing of the hardware, the pricing of the software, and the number of varieties of software available
                         influences which network will emerge in equilibrium. The market outcome diverges from the social
                         optimum in two ways. Lower hardware costs in the future provide a firm with a strategic pricing advantage,
                         while lower software development costs provide a firm with more varieties of software and lower software
                         prices.

Social Systems Simulation Group P.O. Box 6904 San Diego, CA  92166-0904 Roland Werner, Principal Phone/FAX  (619) 660-1603