The nation’s defense industrial base is in serious trouble. Warning lights have been flickering for the past ten years, and now one of them has begun to glow in earnest. Leadership in the design and manufacture of electronic components is moving overseas rapidly. If this trend continues, US military forces will be dependent — within the next decade — on foreign suppliers for critical capabilities they need to maintain their technological superiority.
This is the alarming conclusion reached by a Defense Science Board task force, whose report earlier this year deserves more public attention than it has gotten. The problem revolves around the tiny silicon semiconductor chips that make electronics the dominant technology in modern weapon systems. In the early 1980s, Japan overtook the United States in the semiconductor market and has been pulling further ahead ever since.
The most advance semiconductor today is the one-megabit Dynamic Random Access Memory (DRAM) chip, which stores about a million bits of information on a single wafer. DRAMs were invented in the United States, but the US has been beaten out by the Japanese in the ability to produce these chips in large quantities at very low cost and now supplies less than five percent of world consumption.
It appears that Japan will develop as well as produce the next generation of DRAM chips. If so, the task force says in terse language, the United States in the 1990s will have only two choices: either buy foreign semiconductors or settle for seconds best in its weapons systems. These alternatives assume, of course, that the chips would be available. The task force also observes that “as Japanese firms evolve from the role of merchant semiconductor manufacturers into computer/telecommunications system builders, it would not be an illogical strategic business policy to delay release of the most advanced chips to competitors in the systems market, including those residing in the United States.”
It is true that the Japanese have gotten some of their advantage by means of trade barriers and dumping their products on the world market. But, says the task force, Japan is ahead in semiconductors mainly because of its industrial policies. The Japanese invest more heavily in plants, equipment, and research and development. They work toward long-term goals, effectively integrating the resources of government, industry, and academia. Over time, this gives them the edge in high-volume production, from which nearly all else follows.
The Defense Department is not situated that well to head off the problem. The armed forces, once the primary customers of the semiconductor industry, now buy just three percent of the total quantity produced. The market is driven by commercial demand, not by military considerations.
The best idea the task force could think up was the creation of a semiconductor-manufacturing institute by a consortium of US firms. Its first task, underwritten with substantial Pentagon funding, would be developing the technology to build a sixty-four-megabit DRAM. This is a good idea. If the institute is established, and depending on what else happens in conjunction with its efforts, it could moderate or even correct the drift toward semiconductor dependence. This idea, however, is not a complete solution to the decline of the defense industrial base. That problem is much broader and more complex. Military-industrial first aid will not be enough to set it straight.
The legendary “Arsenal of Democracy” passed into history many years ago. It has been a long time since the US industrial base had a capacity for wartime surge production. As demonstrated by the case of the semiconductors, it cannot always be relied on to meet even peacetime requirements. Between the 1960s and the 1980s, the number of firms doing defense work dropped by more than forty percent. Some, such as foundries, simply folded when environmental requirements become more stringent. Others, especially suppliers of specialized components, shifted their energies to the commercial market, where profits were better and where there was less red tape.
Most of the developed nations of the world are in the process of deindustrialization, and this is one race that the US leads by a furlong. About seventy percent of American workers are employed in the delivery of services rather than goods. Meanwhile, US defense budgets have decreased as a percentage of GNP, so a Pentagon with relatively less money to spend also commands less attention from an industrial base that is itself shrinking in absolute terms.
As if all this were not enough, pressures have built for industry to emphasize short-term profit over long-term development. From the mid-1970s on, a corporation that allowed its dividends to sag became a target for a hostile takeover attempt, with stockholders cheering the sudden boost in share values. In the semiconductor industry, the task force found, “equivalent ownership” of firms turned over completely — meaning that shares of stock traded equaled total stock outstanding — every six to nine months on the average. Stockholders are intolerant of companies that invest for payout five years hence. They want profits in six to nine months. Congress is investigating, but takeovers in US industry increased by a third between 1984 and 1986.
Converging with these events was the notion — which arose from the politics of dissent in the 1960s and soon became pandemic — that the defense firms are a bunch of crooks and profiteers. Few people check deeply enough into the facts to learn that these conclusions are wrong. Industry basking is a popular sport in which no penalties are assessed for fouls. The news media, politicians, and self-styled reformers all play with enthusiasm. Unfortunately, so do some of the more zealous investigators and regulators on the government payroll, who cultivate attitudes that border on contempt. An industry kept in an adversarial, defensive crouch does not serve the nation well as it otherwise might. During the long slide of the industrial base, government policies and rules have often worked to inhibit R&D and productivity investments by defense contractors.
The overall industrial-base problem is so big that no one is able to describe it completely, much less offer a comprehensive solution. The United States should count itself lucky if it can stem the most pressing industrial base issue, the looming dependency on foreign semiconductors.
Even the “domestic” semiconductor industry has moved much of its advanced production capability overseas. Many defense systems already contain components available only from foreign sources. Few military program offices keep records of dependency that look beyond the country of origin of finished devices. The worst part of the problem may be that we do not yet know how bad it is.