The importance of determining optimal policy towards R&D cannot be exaggerated, given the worldwide interest in fostering R&D and given the significant differences between European and U.S. policies towards interfirm cooperation on R&D.' Moreover, the problem is an inherently difficult one because of the complex nature of the R&D process. Since R&D is a component of fixed costs, industries where it is important tend to be concentrated. Hence R&D policy must go hand in hand with competition policy. At the same time, R&D is like any form of investment in that it precedes the production stage. Hence, issues of time consistency and strategic commitment inevitably arise in considering the choice of R&D policy. Finally, R&D by one firm typically leads to spillovers wbich benefit other firms, so that R&D exhibits many of the character- * Department of Economics, University College Dublin, Belfield, Dublin 4, Ireiatid. For helpful discussions and comments, we are grateful to two anonymous referees; to Pedro BarTos, Steve Martin, Massimo Motta, and Damien Neven; and to participants at seminars in Birmingham, St. Andrew's, and Warwick; at a CEPR Workshop on R&D Spillovers in Lausanne; at ERWIT 94 in Caslel Gandolfo, EEA 95 in Prague, the 1996 Conference on Industrial Economics at Universidad Carlos III de Madrid, and a meeting of the Dublin Economics Workshop. This paper was produced as part of the International Economic Performance Programme of the Centre for Economic Performance, London School of Economics, supported by the U.K. ESRC. An earlier version was circulated as CEPR Discussion Paper No. 1243. ' See Alexis Jacquemin (1988) and Stephen Martin (1996). istics of a public good, albeit one that is mostly privately produced. Tbe degree to which such spillovers occur and can be internalized is another crucial influence on the desirable pattem of intervention.