The most important role for government in creating the conditions for commercial innovation is to support universities in their efforts to generate research and provide manpower. The most crucial issue we face is a lack of skilled manpower, a shortage of faculty in universities for training that manpower, and a deteriorating research capability in our great universities because of the shortages of both faculty and modern equipment for instruction and for research. American industry today simply cannot get enough of the people it needs in such fields as microelectronics, artificial intelligence, communications, and computer science. The universities are not turning out enough R85) (research and development) people in these areas, or enough research faculty. There is little that private companies can do about fids. We contribute to the support of universities, but industry will never be able to meet more than a small fraction of university R & D funding needs. Even after a decade of steady increasing industry support for universities, industries provide only about 5 percent of total university R&D funding. Congress is considering additional incentives for industry support of universities, but the fact remains that the primary responsibility for ensuring a strong, healthy academic research system and thereby for providing an adequate supply of research and skilled people must rest with the federal government. There is wide agreement that the federal government should support the universities, and, in fact, federal basic research obligations to universities and colleges, measured in constant dollars, have grown by more than 25 percent over the past three years. But this is only a start in filling the needs. Department of Defense fund lng of basic research, for example, has only in the past two years returned to the level, measured in constant dollars, that it was in 1970. Universities have had to compete with the national laboratories for the Department of Energy’s research dollars. When research is fund at a university, not only does the research get down, but also students are trained, facilities are upgraded, faculty and students get more support, and thereby better faculty and students are attracted. Moreover, the students that go into industry help in the transition of advanced research into concepts for industrial innovation. When the same research is funded at a national laboratory, most of the educational dividends are lost. Universities should not have to compete head on with national laboratories for mission agency funds. Un less the national laboratory will do a substantially better research job, the university should get the funds. The same holds for government funding of research in industry. Those funds that advocates of industrial of policy propose to invest in government-directed industrial R&D would normally be much better spent in universities, unless there is a special reason why an industrial laboratory can do it much, much better. I am not proposing that we simply throw money at universities. We need to be selective. To borrow a phrase from the industrial policy advocates, the government should stress the growth of 'sunrise science and technology. 'Unlike the targeting of sunrise industries, the targeting of-that is, fast moving-areas of research can be done. We can identify these technologies, even if we cannot specify in advance precisely what products or industries they will generate. But we arc not doing this as weft as we can and should. In micro electronics, for example, a study by the Thomas Group, a Silicon Valley consulting firm, concludes that government support of university microelectronics programs totaled only about $100 million between 1980 and 1982. To put that into perspective, the Department of Energy's program expense for just one unproved, highly speculative energy technique, magnetically contained fusion, was $ 295 million in