The Myth of STEM Labor Shortages

(Editor’s note: This is the second in a series of three articles about the push to increase the number of students in science, math, and engineering. The other articles can be found here and here.)

Everybody knows that the best way to get ahead today is to get a college degree.  Even better is to major in one of the STEM (science, technology, engineering, and math) subjects, where the bulk of the jobs of the present and future lie. Politicians, business leaders, and academics all herald the high demand for scientists and engineers.

But they are, for the most part, wrong. The real facts suggest that, in many STEM specialties, there is a labor glut, not a shortage.

That is not to say that the STEM subjects aren’t worthy of study—there are many reasons to do so. And if a talented young person really wants a job in a STEM field, he or she can eventually get one, with a little perseverance.

But there is no urgent need for STEM graduates, at least not in a general sense.

The roots of STEM labor gluts go back over half a century, according to Michael Teitelbaum, a demographer at the Alfred P. Sloan Foundation and a leading authority on this topic. Sputnik, the 1957 Soviet missile launch, created a national concern that we were falling behind in the race for technical superiority. Talk of labor shortages in science and engineering arose, and talk led to action. Beryl Lieff Benderly, a journalist who writes about employment for scientists for Science magazine and other publications, described the result of the national response to Sputnik: “Federal money swiftly poured into science and engineering scholarships and so successfully attracted students that, by the early 1970s, the market for young scientists was flooded.”

The flood grew in the 1980s, after the National Science Foundation (NSF) warned of imminent shortages of scientists. Eventually, the NSF’s predictions turned out to be so off-base that the agency was subjected to an investigation by a House subcommittee in 1995, during which NSF director Neal Lane flatly stated, “there really was no basis to predict a shortage,” according to Teitelbaum and Benderly. (Lane was not involved with the NSF at the time of shortage prediction).

More recently, a PhD. in electrical engineering who follows labor trends in his specialty, Dan Donahoe, wrote for the Institute of Electrical and Electronic Engineers (IEEE) magazine that there has been a “myth of a qualified labor shortage” in his field for a long time. He says that the myth started in the late 1980s, and that the myth continues despite expansion and contraction of the labor market.

The apparent misinformation continues to this day. Microsoft founder Bill Gates has been particularly vocal about supposed shortages of skilled labor in the computer industry.  Microsoft has aggressively used foreign labor to make up for the apparent lack of qualified computer professionals, particularly programmers. According to the employment website, which provides visa statistics based on information compiled from U.S. government agencies, “Microsoft Corporation has filed 33,934 labor condition applications for H1B visa and 10,918 labor certifications for green card since 2001, ranked 1 among all visa sponsors.”

Ironically, Microsoft’s foreign recruitment came on the heels of the dot-com bust that threw huge numbers of American computer professionals out of work shortly after the new millennium.

The simultaneous drop in demand and high levels of foreign recruitment almost assuredly contributed to a subsequent decrease in the number of students choosing computer science as a major. In 2004, there were 59,488 bachelor’s degrees in computer science, according to NSF statistics. Those graduates had entered college just before the crisis of 2001. But five years later, in 2009, the number of graduates in computer science fell to 37,994. It is not too much of a stretch to assume that much of the decline was due to students recognizing that the opportunities in computer science were becoming more limited.

Had there been a shortage of computer professionals instead, the trend would likely have been reversed.

Adding to the problem is universities’ desire to keep academic departments filled with students; when American students, responding to lower salaries and less chance of employment, choose other majors, it appears that colleges simply replace them with foreign students, especially when it comes to graduate students. Sixty-two percent of U.S. graduate students in computer science hold visas from other countries, and 54 percent of graduate students in engineering (according to the NSF). The percentages of foreign graduate students have been on the increase since the 1990s.

The ability of colleges and businesses to replace domestic STEM students and workers from overseas has created a vicious cycle. As foreign workers rush in to the U.S. labor market, the discipline becomes less attractive to domestic students. The failure of students to graduate in STEM disciplines leads to further claims of shortages, followed by more overseas recruitment, making the field even less attractive, and so on.

For instance, in the period between 2001 and 2005—the depths of the bust—only 19,910 net computer-related jobs were added to the U.S. economy, but there were 263,471 new H1B visas for computer-related jobs, according to the Center for Immigration Studies. It’s no wonder that American students fled the discipline, as reflected in the drop in degrees granted between 2000 and 2009 cited above.

Because the influx of foreign students and workers prevents markets from adjusting, temporary shortages that would ordinarily take care of themselves in a few years become labor supply gluts that take decades to sort out.

STEM Jobs 101: Supply and Demand

The vested interests who clamor for more STEM graduates tend to disregard the most basic economic principles, including the laws of supply and demand and the insight that people respond to incentives.

That such principles still actually work in labor markets is on display in the energy industry and in the engineering professions. On the positive side, there has increasing demand for petroleum engineers recently, as the United States steps up development of its domestic production of oil and natural gas.

As a result, the average starting salaries for bachelors–level petroleum engineers doubled from 1997 to 2010, from an already-high $43,674 to an astronomical $86,220, with most of the increase coming since 2005, according to Leonard Lynn of Case Western Reserve, Daniel Kuehn of American University, and Hal Salzman of Rutgers and the director of the J.J. Heldrich Center for Workforce Development (Their study can be accessed here). As salaries have risen, the field has attracted more students. Lynn, Kuehn, and Salzman determined that, between 2002 and 2009, the number of degrees conferred for petroleum engineering in the U.S. rose from roughly 250 to almost 700—more than double in a relatively short time period.

With U.S. students flooding into the field, it should only be a short time before wages level off and even decline a bit—without a huge influx of foreign workers.

Unfortunately, petroleum engineering accounts for only a small percentage of engineering jobs. Plus, the flipside of supply and demand works as well: according to Dan Donohoe, “[T]he United States produces almost four times as many [electrical and electronic] engineers annually as the economy demands.”

For instance, using the federal government’s forecasts, the number of engineering jobs will grow more slowly than the rest of the labor force between 2010 and 2020. The BLS estimated that the total labor force was 153.9 million jobholders in 2010—it is expected to add roughly 20.5 million new jobs by 2020, an increase of 14.3 percent. Engineering jobs—a total of 1.34 million in 2010—will increase by only 11 percent in that period, with slower-than-average growth predicted for ten of the fifteen engineering specialties.

Of the six specialties with the most current jobs—86 percent of total engineering jobs in 2010—only civil engineering is expected to keep pace or exceed the rest of the economy in job growth. Given the construction industry’s ongoing woes, that estimate now seems to be excessively optimistic.

With engineering jobs drying up (or increasingly taken by visa-holders), there has been little to attract more young people to the profession, and fewer are opting to major in engineering. Donahoe produced the following chart which “shows declining student interest in engineering provided by the ACT, Inc. testing services company”:

Donahoe said that the myth has caused the launch of many initiatives to get more students interested in STEM, yet they inevitably fail. Fingers have pointed in many directions to explain these failures, but Donahoe says that they have not pointed out the real reason: “The best way to increase interest in STEM careers is by making certain that STEM careers are actually viable.” In other words, allow demand to catch up with supply.

Where Will the STEM JOBS Be?

So where will the demand be for young American STEM graduates in the future?

According to the BLS, one area is in construction, which requires many civil and structural engineers and architects. But much of this “growth” is merely replacing the huge job losses caused by the housing bust. And the BLS says it won’t happen immediately: “Despite the fast projected growth rate, employment in the industry is not expected to recover to its prerecession level by 2020.”

Furthermore, it seems likely that the BLS projected much of the job growth based on expectations that the government would be funding many infrastructure projects, an idea forwarded by the Obama administration. Given the emerging political climate that favors cost-cutting rather than deficit spending, such an infusion of money into public building projects is far from certain. 

And construction should follow, not lead, the rest of the economy; we are now suffering from an attempt to drive the economy with construction on borrowed money.

Another supposed boom sector is the software industry—it projects a 30 percent increase in “software developer” jobs between 2010 and 2020 (and 22 percent in “systems analyst” jobs).

Yet is hard to see why. Although many important decision-makers still view the high-tech sector as something new with unlimited potential for growth, it is instead a mature industry that has been contracting for over a decade. The Internet is 20 years old; the Microsoft Office software suite is 22. Young people already spend vast amounts of time networking, downloading, and playing video games—how much more can they play in the future? Cloud computing, which moves software off of individual computers and onto the Internet, will account for some increase in jobs, for a while. While there will always be some demand for new software, most software projects today and in the future will be enhancements to or revisions of existing programs, or for small modifications of new hardware and equipment—not the sort of breakthroughs that require massive hiring.

Eventually, the software industry will be forced to do what other industries do as they mature. For instance, the trends in high-tech manufacturing are not promising. Donahoe cites the National Science Foundation’s “Science and Engineering Indicators” that state that the peak for the five main high-tech manufacturing sectors was in 2000. Employment in those industries declined by 687,000 by 2010—28 percent of the total.

Right now, we are at a period of great uncertainty. Predictions of job growth—in any field, should be taken with a grain of salt. We need to let the STEM markets sort themselves out before pushing for more graduates to enter the field. If a young person has a serious interest in one of the STEM subjects, they should pursue it, by all means. But to take a STEM major simply because of the assumed potential for a lucrative job offer may be the wrong way to go.


  • Mark

    Well if you recognize that most H-1Bs are issued under fraudulent, if not illegal premises, then most H-1Bs are in the United States illegally. A H-1B application tainted by fraud means that the H-1B visa therefore issued is not valid.

  • Mark

    Yes, very true. And so much US citizen talent is being destroyed in this fashion. A friend of mine has a PhD in EE, never was able to find a job in the field and basically lives off of his wife now who runs a hair styling business. Hardly an appropriate use of talent, to leave him on the couch while foreign nationals are imported to take the jobs.

  • Mark

    No, employers are not paying “good money” for STEM talent. Most STEM workers earn less than the average police officer.

    As far as people who leave after being trained, the problem there is that the company that does the training simply isn’t paying enough.

    And if you’re hiring engineers without degrees, well, that’s profoundly disrespectful to the legitimate engineers with degrees.

  • Mark

    CS/Math may have higher than average pay, but most of the jobs are located in very high cost of living locales which explains most of the higher pay. If you standardize the locations, STEM workers, particularly in CS/Math, are usually paid less than local police officers.

  • Mark

    The young aren’t getting hired either. Have a look at this piece from Norm Matloff, a UC Davis CS professor:

    “I used to take an informal job survey of graduating seniors every
    year around this time, but had not done so for a few years. So, I tried
    it again last week.

    Of those responding, 46% did not yet have a job, well below the 81%
    figure, but the overall figure is likely considerably higher than 46%,
    for the following reason.

    I got about a 20% response rate, about half of which came only after I
    sent a followup reminder. That’s not too bad, but my experience over
    the years has been that the ones who don’t respond typically don’t have
    a job, and are embarrassed to say so.”