these measures is dificult to establish, being based on assumptions that are not readily tested in objective terms. Nevertheless, one point is fairly clear: As between stimulating industrial R&D by individual firms and stimulating R&D joint ventures of such firms, the joint venture approach appears superior in its likely results. INTRODUCTION The federal government's role in supporting research and development (R&D) is well established. In 1982, a typical year, federal support accounted for 46.7% of all U S . R&D expenditures, about $36.1 billion.' Despite the fact that research universities have a prominent role in the U.S. R&D system, industry is the dominant R&D performer (at least in terms of dollars invested) and nearly half of all federal R&D expenditures is now allocated to the industrial sector. Recent policy initiatives further accentuate the relative importance of the federal government in R&D performance. The R&D tax credit provision of the Economic Recovery Tax Act of 1981 aims at stimulating industrial R&D, and revisions of antitrust legislation are being proposed as an instrument for encouraging additional private sector spending for R&D. Nevertheless, although the interest in increasing private spending on R&D is wide1 shared, there is disagreement about the efficacy of the tax credit and the likely effects of changes in antitrust provision^.^ More basic, there Y * The authors gratefilly acknowledge the helpful comments and guidance of Ray Vernon and anonymous referees.