The Myth of the Science and Engineering Shortage
American students need to improve in math and science—but not because there's a surplus of jobs in those fields.
Everyone knows that the United States has long suffered from widespread shortages in its science and engineering workforce, and that if continued these shortages will cause it to fall behind its major economic competitors. Everyone knows that these workforce shortages are due mainly to the myriad weaknesses of American K-12 education in science and mathematics, which international comparisons of student performance rank as average at best.
Such claims are now well established as conventional wisdom. There is almost no debate in the mainstream. They echo from corporate CEO to corporate CEO, from lobbyist to lobbyist, from editorial writer to editorial writer. But what if what everyone knows is wrong? What if this conventional wisdom is just the same claims ricocheting in an echo chamber?
The truth is that there is little credible evidence of the claimed widespread shortages in the U.S. science and engineering workforce. How can the conventional wisdom be so different from the empirical evidence? There are of course many complexities involved that cannot be addressed here. The key points, though, are these:
Science and engineering occupations are at the leading edge of economic competitiveness in an increasingly globalized world, and science and engineering workforces of sufficient size and quality are essential for any 21st century economy to prosper. These professional workforces also are crucial for addressing challenges such as international security, global climate change, and domestic and global health. While they therefore are of great importance, college graduates employed in science and engineering occupations (as defined by the National Science Foundation) actually comprise only a small fraction of the workforce.
A compelling body of research is now available, from many leading academic researchers and from respected research organizations such as the National Bureau of Economic Research, the RAND Corporation, and the Urban Institute. No one has been able to find any evidence indicating current widespread labor market shortages or hiring difficulties in science and engineering occupations that require bachelors degrees or higher, although some are forecasting high growth in occupations that require post-high school training but not a bachelors degree. All have concluded that U.S. higher education produces far more science and engineering graduates annually than there are S&E job openings—the only disagreement is whether it is 100 percent or 200 percent more. Were there to be a genuine shortage at present, there would be evidence of employers raising wage offers to attract the scientists and engineers they want. But the evidence points in the other direction: Most studies report that real wages in many—but not all—science and engineering occupations have been flat or slow-growing, and unemployment as high or higher than in many comparably-skilled occupations.
Because labor markets in science and engineering differ greatly across fields, industries, and time periods, it is easy to cherry-pick specific specialties that really are in short supply, at least in specific years and locations. But generalizing from these cases to the whole of U.S. science and engineering is perilous. Employment in small but expanding areas of information technology such as social media may be booming, while other larger occupations languish or are increasingly moved offshore. It is true that high-skilled professional occupations almost always experience unemployment rates far lower than those for the rest of the U.S. workforce, but unemployment among scientists and engineers is higher than in other professions such as physicians, dentists, lawyers, and registered nurses, and surprisingly high unemployment rates prevail for recent graduates even in fields with alleged serious “shortages” such as engineering (7.0 percent), computer science (7.8 percent) and information systems (11.7 percent).
Over time, new technologies, price changes, or sharp shifts in the labor market can create rapid rises in demand in a particular occupation. When that happens, the evidence shows that the market seems to adjust reasonably well. Entire occupations that were previously unattractive and declining, such as petroleum engineering in the 1980s and 1990s, have rather suddenly become attractive and high-paid—due to increased energy prices and new technologies for domestic extraction of oil and gas. Others, such as those linked to manufacturing and construction—industries in which well over half of all engineers are employed—have declined over the same period. Surprisingly, some of the largest and most heavily financed scientific fields, such as biomedical research, are among those with the least attractive career prospects, as a recent blue-ribbon advisory committee reported to the Director of the National Institutes of Health. Biomedical Ph.D.s are unusually lengthy and often require additional years of postdoctoral training, yet after completion those with such degrees experience labor market demand and remuneration that are relatively low.
Labor markets for scientists and engineers also differ geographically. Employer demand is far higher in a few hothouse metropolitan areas than in the rest of the country, especially during boom periods. Moreover recruitment of domestic professionals to these regions may be more difficult than in others when would-be hires discover that the remuneration employers are offering does not come close to compensating for far higher housing and other costs. According to the most recent data from the National Association of Realtors, Silicon Valley (metro San Jose) has the highest median house prices in the country, at $775,000—nearly four times higher than the national median.
Far from offering expanding attractive career opportunities, it seems that many, but not all, science and engineering careers are headed in the opposite direction: unstable careers, slow-growing wages, and high risk of jobs moving offshore or being filled by temporary workers from abroad. Recent science Ph.D.s often need to undertake three or more additional years in low-paid and temporary “postdoctoral” positions, but even then only a minority have realistic prospects of landing a coveted tenure-track academic position.
Among college-educated information technology workers under age 30, temporary workers from abroad constitute a large majority. Even in electrical and electronic engineering—an occupation that is right at the heart of high-tech innovation but that also has been heavily outsourced abroad—U.S. employment in 2013 declined to about 300,000, down 35,000 and over 10 percent, from 2012, and down from about 385,000 in 2002. Unemployment rates for electrical engineers rose to a surprisingly high 4.8 percent in 2013.
Claims of workforce shortages in science and engineering are hardly new. Indeed there have been no fewer than five “rounds” of “alarm/boom/bust” cycles since World War II. Each lasted about 10 to 15 years, and was initiated by alarms of “shortages,” followed by policies to increase the supply of scientists and engineers. Unfortunately most were followed by painful busts—mass layoffs, hiring freezes, and funding cuts that inflicted severe damage to careers of both mature professionals and the booming numbers of emerging graduates, while also discouraging new entrants to these fields.
- Round one from the decade immediately following World War II, waning a decade later.
- Round two following the Sputnik launches in 1957 but waning sharply by the late 1960s, leading to a bust of serious magnitude in the 1970s.
- Round three from the 1980s Reagan defense buildup, alarming Federal reports such as “A Nation at Risk” (1983), and new Federal funding for the “war on cancer.” Most of these had waned by the late 1980s, contributing to an ensuing bust in the early 1990s.
- Round four from the mid-1990s, driven by concurrent booms in several high-tech industries (e.g. information technology, internet, telecommunications, biotech), followed by concurrent busts beginning around 2001.
- Round five from the rapid doubling of the National Institutes of Health budget between 1998 and 2003, followed by a bust when subsequent funding flattened.
Each of these rounds was accompanied by excessive claims, and a notable lack of credible evidence. Rounds one through three were motivated by existential Cold War concerns, with advocates focused on expanding the numbers of US students pursuing higher education and careers in science and engineering. As I discovered while researching my book, during rounds four and five, after Cold War security concerns had waned, shortage claimants focused on visa policies that enabled U.S. employers and universities to recruit large numbers of temporary workers and graduate students from countries (especially China and India) that had rapid growth in science and engineering graduates but much lower income levels.
One thing we might reasonably conclude is that over the past six decades there has been no shortage of shortage claims. But what about the present and foreseeable future?
Since 2005 a series of influential reports have been produced by respected organizations and individuals, once again pointing to alarming current (or more commonly “looming”) shortages due to failing K-12 education. Three such reports were published in 2005 alone, by the Council on Competitiveness, by a special committee appointed by the National Research Council, and by a group of 15 business and technology organizations. Were these the opening salvos of the “alarm” stage of another 10-15 year cycle of alarm/boom/bust, the sixth such cycle since World War II? A deep recession with high unemployment has intervened, and in any case we would not be able to know for sure until another 5 or more years have passed.
These publications report correctly that the average performance of American K-12 students is middling in international testing. These data also show that this average performance results from large numbers of both high-performing and low-performing US students. The average national scores reflect both ends of the scale, yet there continues to be a large pool of top science and math students in the U.S. OECD data on “high-performing” students suggests that the U.S. produces about 33 percent of the world total in this category in the sciences, though only about 14 percent in mathematics.
No one should conclude from this that American K-12 science and math education does not need major improvement. Emphatically to the contrary: Every high school graduate should be competent in science and mathematics—essential to success in almost any 21st century occupation and to informed citizenship as well. But there is a big disconnect between this broad educational imperative and the numerically limited scope of the science and engineering workforce.
Editorial writers in respected publications continue to assert that American student interest in these fields is low and declining. Yet according to a recent report from ACT, the college admissions testing service, “student interest in STEM [Science,Technology, Engineering, Mathematics] is high overall,” characteristic of some 48 percent of high school graduates tested in 2013. American high-school students are taking more math and science courses than ever before. Meanwhile UCLA’s respected annual surveys of entering college freshmen show that over the past several years nearly 40 percent have been reporting intentions to major in a STEM subject, not only a large fraction but also a substantial increase from past decades—this percentage was about 32 to 33 percent from 1995 to 2007.
Some of these students do change their minds and complete their degrees in different fields, but others shift into science and engineering majors. As noted earlier, the outcome is that the numbers of science and engineering graduates is at least double those being hired into such occupations each year.
The evidence all points to high levels of student interest, high-performance levels among the students most likely to pursue majors and careers in science and engineering, and large numbers of graduates in these fields.
Ironically the vigorous claims of shortages concern occupations in science and engineering, yet manage to ignore or reject most of the science-based evidence on the subject. The repeated past cycles of “alarm/boom/bust” have misallocated public and private resources by periodically expanding higher education in science and engineering beyond levels for which there were attractive career opportunities. In so doing they produced large unintended costs for those talented students who devoted many years of advanced education to prepare for careers that turned out to be unattractive by the time they graduated, or who later experienced massive layoffs in mid-career with few prospects to be rehired.
Recent forecasts of looming shortages of scientists and engineers may prove to be self-fulfilling prophecies if they result in further declines in the attractiveness of science and engineering careers for talented American students.