HIGH-TECH RESEARCH-AND-DEVELOPMENT SUBSIDIES - ESTIMATING THE EFFECTSOF SEMATECH

Sparked by concerns about their shrinking market share, 14 leading U.S . semiconductor producers, with the financial assistance of the U.S. g overnment in the form of $100 million in annual subsidies, formed a jo int R&D consortium - Sematech - in 1987. Using Compustat data on all U .S. semiconductor firms, we estimate the effects of Sematech on member s' R&D spending, profitability, investment, and productivity. In so do ing we examine two hypotheses: the 'commitment' hypothesis that Semate ch obligates member firms to spend more on high-spillover R&D, and the 'sharing' hypothesis that Sematech reduces duplication of member R&D spending. Whereas the commitment hypothesis provides a rationale for t he government subsidies, the sharing hypothesis does not. We find that Sematech induced members to cut their overall R&D spending on the ord er of $300 million per year, providing support for the sharing hypothe sis.