In the high-stakes semiconductor business, Japan has the Ministry of International Trade and Industry (MITI) and the United States has the Defense Department. At least that is the way it has always been. In Japan, the powerful MITI coordinates research and development. In the United States, the Defense Department has played a similar -- although less formal -- role by funding military research, which helped to spur the growth of the industry and provided spin-off technology with commercial applications.

But now, according to a report due out sometime this fall, the Defense Department is no longer serving as MITI's opposite number. The recent Pentagon drive to create very-high-speed integrated circuits (VHSICs) won't help the U.S. semiconductor industry maintain its leadership position. In fact, the move could possibly hurt the industry.

The problem is that the Pentagon kept the technological advances away from the "dynamic small-merchant segment" that has accounted for much of the industry's success so far, says the report by the Berkeley Roundtable on the International Economy (BRIE). Its conclusion is that the Defense Department can no longer promote commercial competitiveness in the semiconductor industry.

VHSICs (pronounced vee-sicks) are very fast chips that let computers react almost instantaneously to new input. They have wide applications in the military, so the Pentagon wants to start using VHSICs as soon as possible.

Contracts for the program have gone only to large companies, but that isn't the real problem, according to the report from BRIE, a public-policy center at the University of California. BRIE researchers Leslie Brueckner and Michael Borrus argue that ordinarily, small companies benefit by adapting techniques coming out of such a project. But Reagan Administration restrictions on the flow of information from Defense Department projects, the report suggests, may keep many research results secret.

"If only military contractors end up getting access to these innovations, small businesses would suddenly be confronted with a few military contractors that have access to innovations that small companies don't," says Borrus. Small business, crucial to progress in the U.S. semiconductor industry, would be locked out. The industry would be deprived of improvements and new products small companies could have made, leaving it at a competitive disadvantage.

Even without the restrictions, the Pentagon project may not help American business, the report says. Unlike commercial customers, the military is willing to trade off cost savings for earlier delivery of the chips. So companies involved in the project are developing technique that may not be profitable commercially Meanwhile, the Japanese, supported by commercially oriented government programs, may take over the mass market.

"If you're talking about industrial policy, this ain't it," says Borrus. He envisions commercial programs establishing performance standards, similar to government-set military ones, for VHSIC technology. Companies would then bid for money to do research on the project. Once public money was invested, he says, the market could take over.