Innovation Files has moved! For ITIF's quick takes, quips, and commentary on the latest in tech policy, go to

Debunking the myth of a STEM surplus

A new data release by the Census Bureau which claims that only 26 percent of STEM workers end up in STEM fields has seemingly strengthened arguments that America does not face a STEM-worker shortage. The surprising statistic has generated coverage from major news sources (including the USA Today and the Washington Post) which have pounced on the new data as evidence that there is no need to encourage students to study science, technology, engineering, and math. The data, however, is highly misleading and skews the reality of demand for and scarcity of a highly skilled math and engineering workforce.

First, let’s start with the definition of STEM graduates. To the Census Bureau, that means not just individuals with a degree in computers, math, statistics, engineering, biology, or the physical sciences (what the average person thinks of when they STEM)  but also psychology and social sciences like economics and anthropology.

While psychology and social sciences graduates do technically study science in the respect that academic research in these fields attempt to rigorously and empirically tests hypotheses using the scientific method, these fields are a far cry from what readers imagine when they see the label STEM.

Naturally, social science majors do not frequently go into engineering or computer fields, which constitute 80 percent of jobs defined by the Census as STEM, at the same rate that engineering or math majors do. While social science jobs are listed under STEM occupations, they comprise a scant 0.6 percent of total occupations. Additionally, many degrees in life science and psychology prepare students for jobs in health care fields, which aren’t counted in STEM totals. While only 14 percent of life science grads go into STEM fields, an additional 30 percent go into health care. Psychology graduates  follow a  similar pattern.

When looking just at computer scientists, mathematicians, statisticians, physical scientists and engineers, 48 percent go into STEM occupations revolving around computers, engineering, and mathematics, significantly higher than the 26 percent touted by STEM shortage skeptics. Moreover, STEM educated workers comprise 63 percent of jobs in STEM fields that require a college degree.

So where do the other 52 percent go if not to STEM occupations? As it turns out, many of the skills taught in STEM courses are not only transferable to other fields, they are highly coveted. In today’s economy, computer and engineering skills, the ability to work with big data tools, and critical thinking skills gained from studying math, hard sciences, and engineering are in high demand throughout all sectors of the economy. Twenty-one percent of STEM graduates go into business and finance or become managers. Another 10 percent go into sales or office support jobs.  Additionally, 5 percent pursue healthcare occupations and 7 percent become educators to prepare the next generation to excel in STEM fields. Of STEM workers in non-STEM fields, two-thirds reported that their job was related to their degree, a much more accurate indicator of the practicality of a STEM education. In fact, STEM graduates are actually more likely to have jobs related to their majors than non-STEM college graduates.

Thus, the so-called surplus in STEM workers actually reflects an overly narrow definition of STEM occupations and an overly broad definition of STEM majors, while ignoring the increasing demand for STEM skills throughout all sectors of the economy. Such an ill-defined metric is hardly useful for measuring the supply of STEM labor. Better indicators of demand for STEM skills are average wage rates and average unemployment rates for STEM graduates.

Even during the recession, unemployment for STEM graduates hovered around 4 percent (around 3 percent is considered full employment, factoring in structural delays in matching firms and workers). In fact, there were two job openings listed for every unemployed worker with a STEM degree. Moreover, these workers are highly paid, with a median average wage ($73,290 in 2010) that is over twice the national average.

Low unemployment rates and high wages are indicative of a market that demands workers with more STEM skills. Foreign workers with H-1B visas partially satisfy the excess demand, but in most cases firms would hire American labor if it were available. In contrast to popular belief, foreign workers with H-1B visas are actually typically paid more than American laborers, debunking the low-cost labor substitution myth explaining American demand for foreign knowledge workers.

As the global economy becomes more and more competitive, human capital development, particularly in STEM areas, will become ever more important. Not encouraging students to pursue skills that are in high demand today and are likely to be in even greater demand when they graduate is counterproductive to America’s future economic prosperity. In particular, we should be encouraging minorities and women, who are grossly underrepresented in STEM occupations, to pursue STEM educations. However, the real lesson is that Census and the National Science Foundation should remove social sciences and psychology from the definition of STEM jobs and degrees and only include “real” STEM jobs and degrees.

Print Friendly, PDF & Email

  • Nina Mandarin

    IT solutions are essential in every industry today. I found very interesting news hier:

  • Guest

    Hi Nina, addition this site, do you have another one?

  • Martin Energy

    Hi Nina, addition this site, about IT, do you have another one?

  • Ha! In the ITIF study, “Debunking the Top Ten Arguments Against High-Skilled Immigration,” Nagar uses this article (authored by himself) as a “debunking” reference. Citation [5] in the study. (This is the second questionable reference I’ve found in a quick-check (of two cites). Reference [10] does not support the text.)

    Additionally, something Nagar does not seem to understand about unemployment statistics, is that when a person with a STEM degree, has a last position that is not in a STEM occupation, (s)he is not counted as an unemployed STEM professional. For instance, if your last position was as a bartender, you are counted as an unemployed bartender, regardless of your degree or the position you are seeking.

    The ITIF “study” is loosely about STEM, but focuses on IT for support, here is some raw data that does not support a shortage in IT.

    OES Employment Growth 2004-2013 [1]
    — Math and Computer Sci 2004 = 2,915,300
    — Math and Computer Sci 2013 = 3,696,180
    Total Math and Computer Sci Employment Growth = 780,880

    Bachelors and Associate degrees conferred [2]
    Citizens and Permanent Residents: 2004-2013
    Academic Discipline, Broad (standardized)

    — Engineering = 708,298 (70,830 avg. per year)
    — Physical Sciences = 191,492 (19,149 avg. per year)
    — Geosciences = 46,875 (4,688 avg. per year)
    — Math and Computer Sciences = 945,901 (94,590 avg. per year)
    — Science and Engineering Technologies = 1,485,926 (148,593 avg. per year)

    Total = 3,378,492 (337,849 avg. per year)

    [1] Occupational Employment Statistics: OES DATA

    [2] National Center for Education Statistics (NCES) Data Sources
    * Latest data available is 2013

  • roshan.perera

    What you have is a STEM surplus. I have seen dozens of graduates in nanotech, material science who even got PhDs couldn’t find jobs for years. If you look at even a temporary position like Postdoc/Research associate position there are over 100 applications (for just one temporary position). Research funding is at all time low and STEM graduates are at all time high, so my advice is don’t waste your time and money on STEM programs.

    PS: That huge demand in Asia and Middle East for STEM R&D is no longer available.

  • I just completed a data analysis that addresses the concerns you cite: Additionally, many degrees in life science and psychology prepare students for jobs in health care fields, which aren’t counted in STEM totals.

    The BLS gives us a data-set that defines the STEM occupations, NCES provides a crosswalk to the classification of instructional programs (CIP) that are educational paths to those occupations. Because there is no mismatch between the employment data and the educational data, it does not matter if the life sciences are (partially included) — social sciences are not included.

    In 2017, STEM employment growth slowed, temporary foreign workers assumed 110% of that employment growth (118,740). Meanwhile, degree production for citizens and permanent residents in STEM averages about 500,000 per year — almost three times the average STEM employment growth (182,100 yr).

    Additionally, a little known fact about BLS unemployment statistics. If your last job was as a barista, and you are unemployed, you are an unemployed barista, it does not matter if you have a freshly minted PhD in Computer Science. Likewise, if a computer programmer is laid off and has to take a job at WalMart, (s)he is no longer a computer programmer in BLS statistical analysis.