By the time U.S. Industry in 2000 was published, said Dr. Jorgenson, the dozen or so industries it surveyed had regained lost ground or avoided decline altogether. The U.S. semiconductor industry was conspicuous among these and had by then attained a very strong position relative to its counterparts abroad. That industry was the subject of a chapter written by Dr. Mowery and two colleagues. They showed that the industry had come under severe competitive attack in the mid-1980s. By the end of the decade of the 1990s, however, it had returned to international pre-eminence. The speed of that recovery, said Dr. Jorgenson, was “a very important reason for focusing now on that industry.”
Dr. Jorgenson introduced Dr. Kenneth Flamm of the University of Texas as the person who had taken the lead in studying the economic impact of the semiconductor industry. Dr. Flamm had demonstrated “to the satisfaction of economists” that the semiconductor industry had been the key force in the revival of the competitive strength of industries related to information technology. Of the successful industries identified in U.S. Industry in 2000, well over half had been transformed by the use of information technology, which, in turn, depends on developments in semiconductor technology. This, he said, is another reason to try to understand the semiconductor industry and its future contribution to the economy.
Dr. Jorgenson then turned to his own research and that of colleagues at the Federal Reserve Board of Governors and the Brookings Institution. He described a “growing realization” that information technology is a driving force in the U.S. economy, and he said that the revival of the U.S. economy from its current recession depends heavily on the future of the semiconductor industry. He noted that the semiconductor industry was in the midst of its own severe cyclical contraction, which was even more pronounced than the slowdown of the economy. It had experienced at least three, and possibly four, other downturns during the business expansion that began around 1991 and ended at the end of 2000.
During a 10-year period in which the U.S. economy was generally expanding and the unemployment rate falling, the semiconductor industry was subject to severe cyclical downturns, with the present one being the deepest. One assumption about these downturns is that they are related to a special feature of the industry: its exceptionally high rate of technical progress. The growth of the industry periodically exceeds the capacity of the economy to absorb its products2—until this overhang eventually self-corrects in a painful fashion. Hence the industry creates complex challenges for its managers and substantial ripple effects throughout the economy.
This excess in manufacturing capacity results partially from the fact that the increases in manufacturing capacity—units of clean rooms, for example—are quantum in nature. Another factor contributing to excess in manufacturing capacity is that companies put additional capacity in place to capture a greater market share for new products and technologies. Additionally, ramping up a new technology introduces uncertainties—e.g., uncertainty as to yield—that can exacerbate these issues.