Does the U.S. Produce Too Many Scientists?

Editor's Note: Beryl Lieff Benderly, a fellow of the American Associaton for the Advancement of Science, writes about scientific labor force and early career issues in the Science Careers section of Science.

In this rough-draft article, she argues that the scientific labor market is broken, that the U.S. educational system actually produces too many qualified researchers for too few positions, and that a perverse funding structure perpetuates the problem, among other points. We'd like your views on this topic and suggestions on ways to further develop the article. Please use the Comments section at the bottom of the page.
 
Here are a few questions to get you started:
 
After reading the article, do you disagree with the "almost universally accepted" idea that there is a national "technical talent dearth"?
 
What was your reaction to the assertion that the decline of white males in science indicates a drop in the desirability of science careers?
 
Do you think that U.S. education policy should work on improving the science-math performance of the children at the bottom, overwhelmingly from low-income families and racial and ethnic minorities, rather than the performance of all children?
 
How dismal do the science job prospects described here seem in the context of the broader economy?
 
Do you accept as solutions to the "scientist glut" problem that we need to create better-paying staff jobs in labs, reduce low-paying post-doctoral positions and generally restructure the way that the U.S. staffs and funds its academic laboratories?
 
What are the biggest challenges faced by the "American research enterprise--the indispensable engine of national prosperity"? What do you think should be done to make it better?
 
What will happen over the next decade or so as lab space (and grants) start freeing up as the baby boomers who occupy those posts hit retirement?
 
What other perspectives on science career opportunities would you like to see in this article?

For years, Americans have heard blue-ribbon commissions and major industrialists bemoan a shortage of scientists caused by an inadequate education system. A lack of high-tech talent, these critics warn, so threatens the nation’s continued competitiveness that the U.S. must drastically upgrade its K-12 science and math education and import large numbers of technically trained foreigners by promptly raising the current limit on the number of skilled foreigners allowed to enter the country to work in private industry. “We face a critical shortfall of skilled scientists and engineers who can develop new breakthrough technologies,” Microsoft chairman Bill Gates testified to Congress in March 2008.

But many less publicized Americans, including prominent labor economists, disagree. “There is no scientist shortage,” says Harvard University economist Richard Freeman, a leading expert on the academic labor force. The great lack in the American scientific labor market, he and other observers argue, is not top-flight technical talent but attractive career opportunities for the approximately 30,000 scientists and engineers—about 18,000 of them American citizens—who earn PhDs in the U.S. each year.

“People should have a reasonable expectation of being able to practice their science if they’re encouraged to become scientists,” says labor economist Michael Teitelbaum of the Alfred P. Sloan Foundation “It shouldn’t be a guarantee, but they ought to have a reasonable prospect.” But today, however, few young PhDs can get started on the career for which their graduate education purportedly trained them, namely, as faculty members in academic research institutions. Instead, scores of thousands of them spend the years after they earn their doctorates toiling in low-paying, dead-end postdoctoral “training” appointments (called postdocs) in the laboratories of professors, where they ostensibly hone skills they would need to start labs of their own when they become professors. In fact, however, only about 25 percent of those earning American science PhDs will ever land a faculty job that enables them to apply for the competitive grants that support academic research. And even fewer—15 percent by some estimates—will get a post at the kind of research university where the nation’s significant scientific work takes place.

The competition for science faculty jobs is so intense that every advertised opening routinely attracts hundreds of qualified applicants. Most PhDs hired into faculty-level jobs get so-called “soft-money” posts, dependent on the renewal of year-to-year funding rather than the traditional tenure-track positions that offer long-term security. In the information technology field, meanwhile, experienced professionals blame outsourcing and industry’s preference for cheap young foreigners for blighting their careers. The firms using the largest number of H-1B visas, the type of immigration document that admits highly skilled temporary residents to the U.S workforce, are not supposedly talent-starved American technology companies but Indian-owned firms in the business of outsourcing work from American companies to the subcontinent.

Despite these realities, the existence of a technical talent dearth is nonetheless almost “universally accepted” in political circles, where it plays an important role in shaping national policy on science funding, education and immigration, says Ron Hira, assistant professor of public policy at Rochester Institute of Technology. “Almost no one in Washington” recognizes the “glut” of scientists, nor the damage that lack or opportunity is doing to the incentives that formerly attracted many of America’s most gifted young people to seek scientific and engineering careers, he says.

But the real dearth—the lack of clear pathways into careers that could enable today’s generation of gifted young Americans to become the researchers who make tomorrow’s great discoveries—is convincing more and more of the nation’s best students not to seek careers in fields such as law, finance, medicine and other fields that offer much better short- and long-term career prospects instead of dedicating an average of seven years to PhD study plus an additional five years or more of postdoctoral training now considered necessary to compete for an academic career in many scientific fields.

At the same time, however, the U.S. annually admits large numbers of foreign graduate students and postdocs and finds itself increasingly dependent on an inherently unreliable stream of young foreign scientists, mostly in the country on short-term, non-resident visas, to do much of the routine labor that powers American research. The American research enterprise—the indispensable engine of national prosperity and the world’s leading innovation establishment—has therefore become vulnerable, observers say, to conditions beyond its borders and its control. At the same time, experts note that recruiting sufficient amounts of the talent needed for vital defense-oriented scientific and engineering work that requires security clearances has become increasingly difficult.

Reversing these trends will require concerted national action, critics say, but not the steps propounded by some highly publicized reports and industry leaders, such as increasing numbers of graduate fellowships and H-1B visas—steps that critics argue will only make matters worse. Rather, these observers call for changes in the way that the U.S. staffs and funds its academic laboratories in order to restore the incentives that formerly attracted many of the brightest young Americans to seek careers in research and contribute to maintaining the nation’s longstanding scientific and technical preeminence.

One thing that’s not in short supply are scientifically talented American students, whose academic achievements have been increasing rather than declining in recent years. “Students emerging from the oft-criticized K-12 system appear to be studying science and math subjects more and performing better in them, over time,” said Teitelbaum in Congressional testimony in November 2007. “Nor are [they] lagging far behind comparable students in economically competitive countries, as is oft asserted.”  The number of Americans earning PhDs in science and technical fields has risen by 18 percent since 1985, according to the authoritative Scientific and Engineering Indicators 2008 , published by the National Science Board.

But the demographic that historically provided America’s scientists, native-born white men, is down by about 1,000 a year, although minority (mainly Asian-American) and female PhDs are up markedly. But because white men are traditionally the highest-earning segment of the American population and the group with the widest career options, many observers believe that the decline in their numbers indicates a drop in the desirability of science careers. In addition, numerous professors anecdotally report that more and more of their best undergraduates shy away from science graduate programs in favor of fields with more certain returns. The foreign-born segment of the nation’s scientific labor force has meanwhile grown rapidly, especially among postdocs. The total number of postdocs in U.S. labs is unknown, but it may be as high as 90,000, according to Indicators. At least half of them, experts agree, are non-citizens. This tangle of trends, many observers argue, reflects a labor market gone seriously awry.

Arguments for the shortage based on the inadequacy of American education generally begin with the results of standardized tests used in international comparisons.  Average scores for K-12 students in the U.S. never top those lists in either science or math (although they do in both reading and civics). On one widely cited assessment, Trends in International Math and Science Study (TIMSS), which tested American third and eighth graders between 1995 and 2003 and American 12th graders in 1995 and 1999, U.S. students ranked between fifth and 12th in math and science—results bemoaned by many as dangerously deficient.

But a detailed study of students’ performance on TIMSS as well as on the Programme for International Student Assessment (PISA), another widely reported international comparison test, by B. Lindsay Lowell of Georgetown University’s Institute for the Study of International Migration and Hal Salzman of the Urban Institute in Washington, D.C., suggests otherwise. “Their point is that the average performance of U.S. students on these comparative international tests is not a meaningful number,” Teitelbaum says. Far from trailing the developed world in science education, as some claim, “on PISA, the U.S. has more high-scoring kids in science than any other country” and nearly as many in the top math category as top-scoring Japan and Korea, Salzman says.

But crucially from a statistical standpoint, U.S. students are by far the most diverse of any industrialized country, ranging from some of the world’s best-prepared to some of the worst among the developed countries. On tests comparing the U.S., Japan and five Western European countries, for example, white Americans on average substantially outscored the Europeans in math and science and came second to the Japanese. American whites came first in reading by a wide margin. American black and Hispanic students, however, trailed significantly behind all other groups on average.

But scientists are not generally recruited from the average students, Salzman notes, but from those with the top scores, of whom America has large numbers. Compared with the products of Asian secondary schools, American students “are free thinkers,” says Vivek Wadhwa of Duke and Harvard Universities. “They didn’t spend the last 12 years of their lives memorizing books…. They’ve spent the last 12 years dealing with real problems and solving them. [In America], you can walk up to your teacher and tell her that she’s wrong or he’s wrong.” In Asia, he continues, “you wouldn’t dare do that.”

Raising America’s average scores on international comparisons is, therefore, not a matter of repairing a broken educational system that performs poorly overall, as many critiques suggest, but rather of improving the performance of the children at the bottom, overwhelmingly from low-income families and racial and ethnic minorities. This discrepancy , of course, is a vital national need and responsibility, but it does not reflect an overall insufficient supply of able science students.

Nor do American students lose interest in science once they reach college. "The proportion of all bachelor’s degrees awarded in S&E [science and engineering] has been relatively stable over time, as has the proportion of freshmen in an S&E major," Lowell and Salzman state. Over many years, Indicators notes, the U.S. has produced about three times as many science and engineering degree holders as it employs in those fields. In June 2008, for example, the Bulletin of the American Meteorological Association warned that universities are producing far more meteorology graduates than can find work in the field.

Many ascribe the shortage not to failings of America’s schools, but to the structure of its scientific labor market. “No one who has come to the question with an open mind has been able to find any objective data suggesting general ‘shortages’ of scientists and engineers,” according to Teitelbaum. Salaries, for example, have not risen, as would happen in a shortage. But, says Ted Greenwood, also of the Sloan Foundation, “we have made preparation for scientific and increasingly for engineering careers—and in some respects, these careers themselves—…so unpalatable that it is difficult to attract Americans to these fields. But non-Americans, with different incentive structures, flock to the United States and flock to these fields here” because being a student or a postdoc offers admission to the country. “We need to make science and engineering careers and the preparation for these careers more attractive to Americans,” Greenwood adds.

The root of the problem, many believe, is what Teitelbaum calls the “perverse funding structure for science graduate education… a recipe for instability.” Since the 1940s, when the U.S. government began to invest seriously in civilian research, the work has been done largely at the nation’s universities and paid for through competitive, temporary grants awarded to individual professors by federal funding agencies such as the National Institutes of Health and the National Science Foundation. Since then, these agencies have become the major funders of academic research in this country, and, indeed, the world. The National Institutes of Health, which now dispenses more than $28 billion a year, is largest funder of non-military research on the planet.

Through decisions made haphazardly 60 years ago, “we chose as a country to staff our labs primarily with graduate students and postdocs and a few non-tenured staff people, while other countries have permanent ways of staffing their labs,” often with PhD staff scientists in career positions, says Georgia State University economist Paula Stephan, an authority on the academic labor force. Under some of those other systems, research institutions employ many scientists as long-term, career staff members who have professional-level salaries and clear career paths potentially leading to greater responsibility and leadership.

The American approach of temporarily funded labs staffed largely with student and postdoc labor offers several important advantages. It enlists the finest talent at the nation’s great universities in projects that meet national priorities set by the funding agencies or by Congress. It permits flexibility in selecting studies and researchers and the opportunity for rapid changes in direction because the grants are for specific purposes and last only a limited number of years. It elicits the best ideas and best work from highly motivated scientists because it chooses the grantees through a competitive system of merit rankings done by peer committees composed of academic experts in each field who serve as part-time judges. It frees the government from owning the labs and managing their staffs. And it allows federal dollars to do double duty—produce research results and provide education and support for the graduate students and postdoctoral associates who work on the projects in labs run by professors who pay them out of the grants.

 This system produces superb science, but it has several serious drawbacks from the standpoint of recruiting and retaining scientists. First, it makes the funding of any particular lab inherently unstable and dependent on winning repeated grants and renewals, which places individual careers at the mercy of annual competitions. In times of very tight federal budgets, such as the present, this means that many labs, and even many well-established scientific careers, do not survive. Second, it produces not only educational opportunities and research results, but also a constant stream of newly fledged young researchers who need opportunities to start their own careers. “The way that U.S. staffs its labs puts so much pressure on the system to absorb the continual new cohort. And we haven’t had much luck in absorbing it,” says Georgia State’s Stephan.

That’s largely because a scientist can’t compete for a federal grant—the sine qua non of professional recognition as an established investigator—without the backing of a university or other non-profit institution, and universities generally back only researchers who hold faculty positions, and sometimes only those on the tenure track. Scientists write the grant proposals and do the research, but the grant, which often also provides at least part of the professor’s salary, is technically awarded to the university, which administers it and provides the facilities needed to do the research in return for overhead payments. The limiting factor on young scientists’ abilities to start academic research careers is thus the number of available faculty positions, which over recent decades has fallen farther and farther behind the number of scientists the system is producing.

Despite a longstanding dismal job market in academic science, however, departments continue to recruit graduate students and postdocs because they need skilled and inexpensive labor to do the work promised in professors’ grant proposals. Doctoral-level researchers must receive the “trainee” wages paid to postdocs—generally about $40,000 a year for 60 to 80 hours a week with no job security or promotion opportunities. But paying postdocs a true professional wage would mean many fewer highly skilled hands, fewer publications and less chance of winning a grant renewal.

This dynamic creates distorted incentives, an artificial sense of shortage and a vicious circle. From the standpoint of a department chairman, Teitelbaum says, “you’ve got this research funding [that] will finance 15 graduate research assistants and 10 postdocs and your department and your faculty are committed to doing the research because you won the grants, but there aren’t enough people applying to be graduate students and postdocs from the U.S. From your perspective, that could be deemed to be shortage.” But, he emphasizes, “the demand is inside the institution, it’s not in the labor market.” Faculty members intent on getting the research done are “not thinking about…whether there’s post-university demand for people who have gotten PhDs or done postdocs.”

If Americans are not available to staff a lab, university professors have the option of foreign talent, as graduate students admitted to the country on student visas, or postdocs, admitted under what Teitelbaum calls the “essentially unlimited numbers of H-1B visas” are available to universities because of an exception available to non-profit institutions.

 Some senior academic scientists have told Teitelbaum they are “very worried” about the fact that the supply of scientists that this country’s universities train is thus totally unrelated to the demand for researchers in the market for career positions, but they find it “difficult to be open about it because it's very threatening to the structure by which research is done,” Teitelbaum says. “ Who’s going to actually do the bench research?” Other professors, educated decades ago, “finished their doctorates, in a very tight labor market where they could get a tenure track position or have several offered them right off the bat,” Teitelbaum continues. “So they have only positive views of their experience. And they might think that people today are just complainers or whiners and [so] just get on with it. If you’re good enough you’ll do fine. That would be a fairly typical position.”

Until the late 1960s, the majority of young PhDs actually did land faculty jobs. They typically finished a science doctorate in about four years and moved on, in their late twenties, to a faculty post and a lab of their own. Only a minority spent an optional year or two as a postdoc, either to pursue a particularly promising line of research or to boost their prospects of a job at a top institution by polishing their publication record with a particularly distinguished mentor. But the Sputnik-inspired increase in research funding and graduate fellowships that continued into the late 1960s and1970s soon began producing more PhDs each year than the nation’s colleges and universities could readily absorb. More and more PhDs, educated in a culture that has long viewed—and, in many places, still views—positions outside the academy not as valid career options for serious scientists but as “alternative employment” at best and “going over to the dark side” at worst, began accepting postdoc positions in the belief that additional publications would improve their chance to land that coveted faculty post.

By the late 1970s, Personnel Needs for Biomedical and Behavioral Research, an authoritative report issued every two years by the Institute of Medicine of the National Academies was expressing alarm at the increase in both the number of postdocs and the length of time they spent as “trainees.” The buildup, as they feared, has continued for three decades, with the growing postdoc pool—enhanced by increasing numbers of foreign PhDs--serving to disguise unemployment among young aspirants to university careers. In many fields, five or more years of postdoc “training” has become the norm, while the percentage of PhDs who landed academic positions continues to drop.

As a result, the average age at which the minority of young scientists who do actually land faculty jobs get to launch independent research careers by winning their first competitive grant has risen to 42. At that age, scientists of previous generations, such as Albert Einstein, Marshall Nirenberg and Thomas Cech, were winning their Nobel Prizes for work done in their twenties.

Spot shortages may exist in certain limited fields, especially those that are new or that require citizenship for security clearance. But in general, writes Harvard’s Freeman, “the job market for young scientists and engineers has worsened…relative to… many other high level occupations, which discourages US students…[but] the rewards are sufficient to attract large immigrant flows, particularly from less developed countries,” in a study published by National Bureau of Economic Research. After a small drop following the terrorist attacks of 2001, the number of foreign scientists arriving has rebounded. “The supply of U.S.-born/residents, particularly men, to science and engineering appears to be more responsive to labor market conditions than the supply of the foreign born,” Freeman continues. “Science and engineering careers may be the only way for many talented foreign-born persons to enter the U.S. job market.”

Recruiting abroad "benefits the country by tapping a large and relatively inexpensive pool of talent at the cost of reduced incentives for native-born individuals to go into science and engineering,” he writes. His Harvard economics colleague, George Borjas, for example, has demonstrated that inflows of foreign students and scientists do, indeed, depress opportunities and incomes for both Americans and foreigners. Critics argue that the H-1B, furthermore, also depresses pay because under its terms, the visa belongs to the employer, not the worker, who therefore cannot leave the job that provides the permission to be in the country.

This “pyramid paradigm can’t continue forever,” says Susan Gerbi, chair of molecular biology at Brown University and a longtime critic of the current method of staffing American labs. “We need to have solutions for some new steady-state model” that will make scientific careers once again attractive to the ablest young Americans by providing an appealing and reliable return on the investment of time, money and opportunity that they make in their education.

Gerbi, Stephan and others believe that the U.S. needs to establish “non-replicating” research organizations with many fewer temporary student and trainee lab workers and many more permanent career staff scientists. Such prominent institutions as the Max Planck Institute in Germany, where Gerbi worked, and the Howard Hughes Medical Institution’s new Janelia Farms Research Campus in Ashburn, Virginia, follow versions of this model.

Freeman has another idea: The way to attract young American talent is by providing much higher incomes for graduate students and postdocs, but with the explicit understanding that, as in such other highly prestigious, highly competitive fields as acting, music and professional sports, only a few of those who enter the competition will win the jackpot of a successful long-term career. People will enter knowing, Freeman proposes, that when their support ends in a few years, the also-rans must leave academic research and move on to other types of work, preferably using their scientific knowledge, in industry, government, patent law, science policy, high-school or community college teaching or other fields. Yet another group of observers, including University of Michigan engineering professor William Kauffman, believe that the way to attract bright and clearance-worthy young Americans to train for careers in national security-related research may be to seek them out as early as high school and offer them attractive scholarships that require a commitment—similar to that of ROTC scholarships—to work in the field after they receive their degrees.

Whatever model or models the nation chooses, many observers believe that the existing system of research by professors who constantly produce large numbers of scientists unlikely to achieve their career aspirations is near collapse. The real crisis in American science education is not young Americans’ inability to learn, or the schools’ inability to teach, but a distorted job market’s inability to provide them careers worthy of their abilities.