Chapter 1 examines the positive and normative effects of four different R&D organization modes---R&D competition, R&D cartelization, RJV competition, and RJV cartelization---in a tournament model of R&D. The results indicate that the highest-ranked R&D organization mode for level of effective investment does not always match that for level of social welfare. In addition, the R&D organization mode leading to the highest profits is always the same as that yielding the highest social welfare. These results therefore suggest that a policy goal of fostering R&D might be a wrong one when R&D competition takes the form of a tournament, and more importantly any government policy towards interfirm cooperation on R&D would be redundant and unnecessary.; Chapter 2 examines government R&D policy toward the export market when international R&D rivalry among the oligopolistic firms displays the features of a tournament. Apart from the nature of product-market competition and the relative number of home and foreign firms, I found that the features peculiar to the tournament model of R&D also emerge as key determinants of policy: the cost structures of R&D, the form of technological uncertainty, the magnitudes of innovation, and the patent policy in the import country. Under a variety of circumstances, optimal R&D policy for a given mode of oligopolistic competition (Bertrand and Cournot) generated by the nontournament model can be qualitatively different from that generated by the tournament model.; Chapter 3 compares incentives and efficiency under the tort system (the comparative negligence rule) versus the various no-fault regimes in automobile accident law. The main results are the following: (1) Under pure no-fault, drivers have an incentive to use either the economically efficient level of care or less, but never more. (2) Under pure tort system, drivers may use the economically efficient level of care, or more or less. (3) The mixed no-fault system sets up identical incentives as the pure tort system when the threshold level is zero and sets up identical incentives as the pure no-fault system when the threshold level is infinite. (4) No single system always dominates the others on efficiency grounds; all are socially preferred under different sets of parameter values. (5) The simulation results show that pure no-fault or the mixed no-fault system with a high threshold for opting out are preferred on efficiency grounds under the widest range of parameter values.
展开▼