PETER IS THE AUTHOR OF THE BOOK, "TO BOLDLY GO: A PRACTICAL CAREER GUIDE FOR SCIENTISTS" 
PREVIOUS COLUMNS 
Is it just me, or have you noticed an increase in the number of articles, books, and programs about trends and forecasts of the future? These days, with a presidency in crisis and a world economy teetering on the brink of collapse, talking about the future may be much more soothing than fretting about the present. With the millennium right around the corner, prognostication is in! Unable to resist this tide of fortune telling, I have spent the past few weeks doing my own trend analysis about science employment.
You would think that scientists would be rather shrewd predictors of the future. After all, the future is what we work on every day. But if you look at the track record, you will find that predicting the future of science employment is a perilous process. Past prognostications have turned out to be wildly in error, with terrible consequences for young scientists.
The most notable example of erroneous prediction is the 1990 one made by Richard Atkinson, president of the University of California, on "a looming shortfall in Ph.D. supply." This particular oracular statement led policy-makers and graduate departments to ramp up Ph.D. production. At the time, many young scientists detected a mismatch between Atkinson's predictions and the realities of the job market. However, it took the "establishment" several years to acknowledge the error. For example, the National Research Council recommendation for tighter controls on the supply of life science Ph.D.s  earlier this month is the first official endorsement of a means for rectifying the problem.
Does this poor track record of prediction mean that forecasting future demand for Ph.D.s is impossible? Not necessarily. The first rule of prognostication is the linear extrapolation. In other words, if you spot a trend in the recent past, chances are that the trend will continue in the future.
As it turns out, Atkinson's 1990 predictions assumed several incorrect changes in the trends of Ph.D. production and job supply. For example, Atkinson assumed that the number of foreign-born Ph.D. seekers would hold constant, despite clear trends of a steady increase in the number and proportion of noncitizen Ph.D. students. Atkinson also assumed an increase in the rate of retirement from academia, whereas all the prevailing trends showed that people are remaining in the workforce longer. These assumptions contributed to the gap between his predictions and the actual job market.
With this cautionary tale in mind, let us look at some of the major forces that shape and control the employment of scientists.
Trends in Federal Support for R&D
As you know, money is the mother's milk of science. Practically every measure of growth in science (including number of Ph.D.s and number of publications) is highly correlated to the amount of research funding. Most of the money spent in the United States is in the areas of applied R&D, which is why there is so much growth in engineering and other applied disciplines. But basic research, the land most academic Ph.D.s and researchers inhabit, represents only 15% of the total amount of R&D spending. This is significant because levels of science employment correlate best with levels of basic science support.
For years, AAAS has amassed figures and data on trends in R&D funding  in the United States and abroad. Their data show that the federal government--the major funder of basic research--has sustained a fairly constant level of R&D spending over the past 25 years , with nearly all the growth in total R&D spending in the industrial sector. Industrial investments in R&D ebb and flow depending on the health of the economy and particular sectors. Relatively wealthy sectors, such as information technology and biotechnology, spend proportionately more on research than sectors that must compete for available funding and tend to shave their R&D investments.
Restricting ourselves to the federal government's R&D portfolio, we can see other trends. The proportion of defense-related R&D has fallen substantially since 1990 .* This is due not only to the end of the Cold War but also to an increasing reliance on "commercial off the shelf" technology in new defense systems. Only a major regional conflict involving the United States is likely to reverse this trend. In the nondefense portion of federal R&D spending, only one field of research has shown steady increases in funding, year in and year out: health .
Other areas, such as space and energy, experienced fluctuations in funding as priorities shifted and political crises passed. With the American electorate graying and the government supporting huge health care costs, we can expect the proportion of R&D funding that goes toward health only to increase in the future.
In the last year, several members of Congress have called for a "doubling" of funding for science over the next 5 years. Relative to times past, these are indeed good times for science funding. Science seems to enjoy popular bipartisan support these days, and many members of Congress believe that the federal investment in science will reap substantial economic rewards. However, before any young scientists start banking on a large increase in funding for science, let me caution you that we have heard these words before. Congressional calls for more funding are just that: recommendations. But when it comes to allocating the shrinking wedge of "discretionary spending" (federal dollars that are not already committed to Social Security and other entitlement programs), science has to compete with all those other hungry mouths: education, transportation, housing, etc. Although science may pay out big dividends in the long term, other investments pay out far more handsomely in the "short term," a.k.a. getting reelected to Congress. Because it takes a long time to build up the successful equivalent of a Silicon Valley or a Research Triangle Park, I would be surprised if real funding for science increases by more than 10% to 15% in the next 5 years. Most of that money will probably go into health-related research. So for non-life science research, the expansionary era of science has truly come to an end.
As I mentioned above, industrial funding of R&D is growing as a whole. However, the bulk of that investment is in the "D" and not the "R"! A number of economists, science-policy experts, and government leaders have noted a shift in industrial research away from long-term basic science and toward applied, nearer term research. Many large industrial laboratories, such as Bell Labs, have been dismantled or restructured, and the era of basic "curiosity-driven" research appears to be over in nearly all of them. Many in the science community have bemoaned the relentless pursuit of the short term in industrial research and worry that breakthrough technologies of the future may fail to emerge in such an environment.
However, along with the dismantling of their in-house basic research divisions, many companies and industrial sectors are forging stronger ties with universities--the known repositories of basic science and the source of new scientists. Companies are finding it more profitable and reliable to scour the world for breakthrough technologies in universities, smaller companies, and national laboratories and license those technologies rather than rely on their own research staff to produce the breakthroughs they need.
So the trend away from big, centralized industrial labs is less a retreat from long-term research and more an outsourcing of research. Whereas industrial R&D was once vertically integrated, moving from idea to product under one roof, it is now becoming more distributed among numerous players.
Overall, we are seeing a rise in the mobility of the American workforce. People not only are switching jobs more often but are moving from field to field more frequently. This appears to be happening in science as well. Although tenured scientists in academia or government are more or less immune, they represent the minority. The bulk of science Ph.D.s live and work in industry, where economic cycles have resulted in downsizing and layoffs, followed by recoveries and more hiring. In academia, an increasing number of young scientists are finding employment only in temporary positions, which creates greater mobility. For many young scientists, mobility is becoming a way of life. Although many of them bemoan this situation, there are no indications on the horizon that this trend will change.
New Fields of Science
New scientific discoveries and synergies are creating a dizzying assortment of interdisciplinary fields. The 19th century structure of science maintained by the academic establishment--math, chemistry, physics, and biology--maps less well onto the new landscape of science. New fields, such as bioinformatics , biomaterials , and combinatorial chemistry, require core training in more than one traditional scientific discipline--training that most of academia is still unwilling or unable to supply. We are seeing more and more Ph.D.s trained in an overpopulated field switch to an entirely new discipline, often with intellectually catalytic results. In the hot new field of micro-electromechanical systems (MEMS) technology, a newly minted B.S. inquired whether a Ph.D. in electrical engineering, physics, or biochemistry would be more important. Her prospective summer employer, a bioassay MEMS start-up company, replied, "All three."
Because basic science funding in the United States will probably remain flat relative to overall economic growth, we cannot expect much expansion in the number of research jobs for Ph.D.s. Research related to health care may still have room for growth, but the environment is already very densely populated by Ph.D.s and postdocs.
Growth in new fields and applied areas will probably continue. Engineers, and scientists who can cross over to engineering, will continue to fare well in the job market. Academia will experience growth mostly in areas of teaching and teacher education but not in research. The number of scientists employed by the government will remain steady relative to the size of the government as a whole. However, the Department of Defense (DOD) may continue to downsize as they contract their R&D needs to external companies, as industry has done. Research in national security-sensitive areas outside DOD, open only to U.S. citizen Ph.D.s, may expand. Industrial R&D will continue to be contracted out to universities, smaller companies, and start-ups.
Scientists in industry will encounter mobility, both elected and forced, at about the same rate as is currently experienced. Globalization of the world's economy may open up more opportunities for bilingual Ph.D.s. and noncitizen Ph.D.s. Because people are remaining in the workforce longer in life, and mandatory retirement is ending, it is unlikely that we will see a wave of retirements as predicted by Atkinson and others. The best way to summarize the Ph.D. job supply is as a closed system where Ph.D.s are finding their first permanent position later in life and hanging onto it longer.
Because Ph.D.s are the most cost-effective means of producing research, and because university-industry R&D collaborations are likely to grow, I expect Ph.D. production to continue to rise. Academia is unlikely ever to adopt any significant controls on Ph.D. production. The United States will remain the world's main supplier of Ph.D. education. And because of the 6-plus-year time lag between enrollment and graduation, it is virtually certain that a graduating Ph.D. will encounter a job market significantly different from the one he or she left on entering graduate school.
The Bottom Line: What Can Young Scientists Do?
Here are a few things I would recommend:
Recognize that the "traditional" model (e.g., research in academia or the equivalent) is a closed system. There WILL be academic job opportunities in the future, but they will continue to be scarce and highly competitive. Ph.D.s trained to carry out basic research in only one field will have the fewest options, while cross-disciplinary Ph.D.s will have more choices. The biggest challenges facing future Ph.D.s will be in learning to balance breadth and depth to maximize opportunity. Public-private sector research collaborations will grow. Scientists with experience in both academia and industry stand to benefit the most. Research into new cross-disciplinary fields will be one of the few growth areas. But here the landscape changes fast, and what is hot now will not be hot in 4 years. Scientists who can move into new areas without returning to grad school for retooling will have an advantage. Personal relationships and work experience will continue to be the primary means by which Ph.D.s find jobs. The more young Ph.D.s can develop a network and valuable work experience, the more options they will have. What other trends do you see, and what do you think young scientists should do about them?
Recognize that the "traditional" model (e.g., research in academia or the equivalent) is a closed system. There WILL be academic job opportunities in the future, but they will continue to be scarce and highly competitive.
Ph.D.s trained to carry out basic research in only one field will have the fewest options, while cross-disciplinary Ph.D.s will have more choices. The biggest challenges facing future Ph.D.s will be in learning to balance breadth and depth to maximize opportunity.
Public-private sector research collaborations will grow. Scientists with experience in both academia and industry stand to benefit the most.
Research into new cross-disciplinary fields will be one of the few growth areas. But here the landscape changes fast, and what is hot now will not be hot in 4 years. Scientists who can move into new areas without returning to grad school for retooling will have an advantage.
Personal relationships and work experience will continue to be the primary means by which Ph.D.s find jobs. The more young Ph.D.s can develop a network and valuable work experience, the more options they will have.
What other trends do you see, and what do you think young scientists should do about them?