Ever wonder why the UC does not automatically waive tuition and fees for graduate students who are supported by assistantships, in contrast to the policy of most other major universities? It all stems from a California legislator's displeasure, years ago, with the fact that UC staffs so many of its lower-division courses with teaching assistants rather than faculty.

Years later, this issue has reached a boiling point, as faculty find it difficult to pay their foreign students' nonresident tuition out of their grant funds. Yet proposals for solving this problem miss the point: Even if the tuition problem were solved, we cannot continue to rely on foreign students to support graduate programs in the future.

There has been a sharp decrease in applications by foreign students to American universities nationwide in the last few years. The explanations typically offered are post-9/11 visa restrictions and the like. Yet more careful analyses have shown that the biggest factor is simply economic:

Professional opportunities are declining in the U.S., while they are expanding in the students' home countries. It is just more attractive for them to stay home.Thus, the nonresident tuition problem is on its to way to becoming moot. Sooner or later (likely sooner), our universities will need to turn to U.S. citizens and permanent residents as the primary source of their graduate students.

But why don't more domestic students currently pursue graduate study in science and engineering? The conventional answer to this question is that the American educational system just isn't producing students who are skilled at science and math. They note that the U.S. is outscored by a number of Asian countries in the Trends in International Mathematics and Science Study.

Yet this is highly misleading. California is still grappling with socioeconomic issues, but much of America is doing very well. For example, on the eighth-grade science test, if Colorado, Connecticut, Iowa, Minnesota and 10 other states had been treated as nations separate from the U.S., each of them would have ranked second in the world, surpassed only by Singapore.

And anyone who thinks that we have a "shortage" of Americans who developed good math and science skills in school should talk to all those doctorate-holders in physics who never were able to secure work in the field.

Problem is economic

Instead, the real problem is again economic: It just doesn't pay for most domestic students to pursue a doctorate in science and engineering. The National Research Council found, for example, that a domestic student incurs a net lifetime loss in earnings if he/she pursues a doctorate. The higher doctoral salary does not ever make up for the industry-level income one forgoes during the five or more years of graduate school.

Indeed, the premium paid for a doctorate in science and engineering is surprisingly small. Doctorates in computer science, for instance, make an average of only 38.7 percent more than those having just a bachelor's degree, compared with a salary differential of 116 percent in economics and 150 percent in political science. Average professor salaries show a similar trend: law, $109,478; business, $79,931; biological and biomedical sciences, $63,988; mathematics, $61,761. This, too, has interesting historical roots:

A 1989 National Science Foundation internal report argued a need to limit growth in doctoral salaries in science and engineering, and proposed as a solution bringing in more foreign students and scholars. It recognized the negative impact this would have on domestic student enrollment: "(If) doctoral studies are failing to appeal to...the best citizen baccalaureates, then a key issue is pay. The relatively modest salary premium for acquiring a (science and engineering) doctorate may be too low to attract a number of able potential graduate students."

Increase stipends

One major remedy our universities could institute would be to substantially increase the size of assistantship stipends, say to 60 percent of the level of starting pay for new bachelor's graduates in industry. This would reduce the long-term financial penalty incurred by going to graduate school, and more importantly from a psychological point view, would greatly reduce the immediate opportunity cost.

Yet clearly the core of the solution involves doctoral salaries. If industry wants to keep its research and development work in the U.S. — the jury is still out on this issue — it will need to set doctoral salaries high enough to make graduate study a good financial investment. If it fails to do so, "big science" research in U.S. universities will become untenable.

The U.S. is often accused of being shortsighted, not planning for the long term. Well, here's a case in point. A permanent decline in foreign student interest in U.S. universities is inevitable. The nonresident tuition issue is an immediate problem to be solved, but if it distracts attention from the broader problem looming in the not-so-distant future, academia will have made one of its biggest blunders in history.

Norm Matloff is a professor of computer science.