May 2015

STEM crisis or STEM surplus? Yes and yes

The last decade has seen considerable concern regarding a shortage of science, technology, engineering, and mathematics (STEM) workers to meet the demands of the labor market. At the same time, many experts have presented evidence of a STEM worker surplus. A comprehensive literature review, in conjunction with employment statistics, newspaper articles, and our own interviews with company recruiters, reveals a significant heterogeneity in the STEM labor market: the academic sector is generally oversupplied, while the government sector and private industry have shortages in specific areas.


Economic projections point to a need for approximately 1 million more STEM professionals than the U.S. will produce at the current rate over the next decade if the country is to retain its historical preeminence in science and technology.—President’s Council of Advisors on Science and Technology1

Unemployment rates within STEM fields…are often higher than they’ve been in years—a sign that there is a shortage of jobs, not workers.—Michael Anft2

Over the past decade, there has been substantial concern regarding the adequacy of the science, technology, engineering, and math (STEM) workforce. Opposing sides paint a polarizing picture: Is there a “STEM crisis” or a “STEM surplus”? Our answer is that there are both.

STEM covers a diverse array of occupations, from mathematicians to biomedical researchers, and at degree levels from bachelor to Ph.D. Some occupations have a shortage of qualified talent, such as nuclear and electrical engineering Ph.D.’s who are U.S. citizens; in other areas, such as biology Ph.D.’s aiming to become professors, there is a surplus. Although many studies have examined the science and engineering workforce in the aggregate,3 little analysis has been aimed at identifying specific areas of STEM worker shortage or surplus. Using a “taxicab queuing model” as a framing metaphor, this article examines the heterogeneous nature of STEM occupations by studying distinct STEM disciplines and employment sectors on the basis of current literature and statistical data, as well as anecdotal evidence from newspapers.4 To augment our findings, we interviewed company recruiters from a wide range of industries in order to gauge the ability of employers to fill open positions. We evaluate these interviews by means of labor market data and scholarly work so as to understand better, from a recruiter’s perspective, the hiring needs of employers and the hiring difficulties encountered by STEM workers.

The ongoing STEM debate. Depending on the definition, the size of the STEM workforce can range from 5 percent to 20 percent of all U.S. workers. Although fields such as computer programming and mechanical engineering are generally considered STEM fields, there is less consensus on areas such as medicine, architecture, science education, social sciences, and blue-collar manufacturing work. In this article, “STEM” refers to the science, engineering, mathematics, and information technology domain detailed by the Standard Occupation Classification Policy Committee, but excluding managerial and sales occupations. Under this definition, postsecondary teachers in STEM fields and lab technicians are considered STEM workers, but workers in skilled trades, such as machinists, are not. Our analysis focuses on graduates with postsecondary education within this STEM domain.

Numerous reports detail the growing concern of policymakers and industry leaders regarding a shortage in the STEM workforce believed necessary to sustain the U.S. innovation enterprise, global competitiveness, and national security.5 Most notable is the National Academies’ report Rising Above the Gathering Storm, which called for improvements in kindergarten through 12th-grade science and mathematics education and increasing the attractiveness of higher education, among other recommendations.6 The report highlighted troubling issues in a number of areas: low STEM retention rates, a relative decline in the number of U.S. citizens enrolled in science and engineering graduate school, and lower percentages of STEM graduates than those of other developed countries. These sentiments were echoed in a 2012 report by the U.S. Congress Joint Economic Committee which stated that the current STEM workforce was falling short of demand in both STEM and non-STEM occupations.7 According to the President’s Council of Advisors on Science and Technology, the United States would need to increase its yearly production of undergraduate STEM degrees by 34 percent over current rates to match the demand forecast for STEM professionals.8

There are, however, many who hold a different view. For example, Michael S. Teitelbaum, vice president of the Sloan Foundation, opined that there are no general shortages of scientists and engineers.9 He went even further, to state that there is evidence suggesting surpluses: there are significantly more science and engineering graduates in the United States than attractive positions available in the workforce. Similarly, B. Lindsay Lowell and Harold Salzman have pointed to the disproportionate percentage of bachelor’s degree STEM holders not employed in STEM occupations.10

Looking at the STEM labor market, Salzman and colleagues concluded that, for every two students graduating with a U.S. STEM degree, only one is employed in STEM and that 32 percent of computer science graduates not employed in information technology attributed their situation to a lack of available jobs.11 In 2014, the U.S. Census Bureau reported that 74 percent of those who have a bachelor’s degree in a STEM major are not employed in STEM occupations.12

Taxicab queuing metaphor

The taxicab queuing problem was first documented in the literature by David George Kendall.13 According to the taxicab queuing metaphor, each taxi–passenger system represents a narrow segment of the STEM employment system. Employers or job positions can be thought of as a finite number of taxicabs, and STEM workers can be thought of as a stream of would-be passengers. We have employers searching for employees, analogous to a queue of taxis waiting for passengers, and another queue of STEM workers searching for jobs, similar to how passengers wait for taxis. If the number of employers searching for employees is greater than the number of STEM workers, we have a queue of taxis, which manifests itself in the real world as a STEM shortage. If the number of STEM workers is greater than the number of employers, we have a queue of STEM workers, meaning that there is a STEM surplus. If the number of employers and the number of STEM workers are equal, we have a momentary match between supply and demand and there is no queue.

This queuing theory framework provides a novel approach to looking at the STEM labor market and the STEM crisis-versus-surplus conundrum. The demand and supply of STEM workers vary by market and location in much the same way that the demand and supply of taxicabs and passengers do. Just as there are separate lines for taxicabs that accept credit cards versus ones that do not, there are distinct lines for each type of STEM occupation. The demand for workers with doctorates in mechanical engineering is different from the demand for those with bachelor’s degrees in mechanical engineering, and the supply of workers with doctorates in the biomedical sciences is different from the supply of those with doctorates in physics. There are also spatial differences. A queue of waiting taxis may be a common sight at an airport, but outside a hotel it may be more common to see a queue of waiting passengers. Analogously, the demand for petroleum engineers in Texas is different from the demand for petroleum engineers in Massachusetts. The upshot is that there may not be a STEM “crisis” in all job categories, but instead just in select ones at certain degree levels and in certain locations.

This model also captures the probabilistic nature of supply-and-demand markets. The times at which both employers and STEM workers enter the job market are uncertain. A job segment that traditionally has a shortage of workers may at some times have a surplus and vice versa. Thus, it is probably far more accurate to state that, within STEM job categories, there is a “crisis” or a “surplus” depending on the circumstances at the time the categories are investigated.

Methodology and data

The STEM supply-and-demand dynamics involve many actors: students, current STEM workers, educational institutions, government, and the private sector. Depending on the STEM segment, changes in each of the actors influence the market to varying degrees. Detailed data on STEM labor markets tend to be sparse. On the supply side, underreporting surpluses is a problem: the reported unemployment rate of STEM graduates is consistently low, but does not reflect those who are underemployed or have switched fields. On the demand side, there is little available data on job openings in the aggregate for various STEM job segments.

To analyze the STEM labor market, we used an indepth literature review of available data sources in conjunction with other sources, such as newspaper articles. To obtain firsthand data, we also interviewed talent recruiters from a wide variety of organizations, including government contractors, media companies, information technology companies, research institutes, startups, and consulting agencies. Because of the small sample size (n = 18), the interviews may be limited in generalizability. Hence, interview results are included only when they supplement the literature or fill gaps in it. Our objective is to highlight the heterogeneity of the demand for and supply of STEM workers, rather than paint a complete picture of supply and demand across all STEM job segments.

Literature survey and results

As intimated from the outset, the literature on the supply and demand of STEM workers is bipolar, with one side proclaiming an impending STEM crisis and the other side asserting a STEM surplus. To understand this conundrum better, we examine the STEM market at a deeper level. By segmenting the STEM labor market into different disciplines, sectors, and skill levels, we find that there is considerable heterogeneity in the supply and demand of workers. Our analysis of the STEM labor market is broken down into three main employment sectors—academia, government, and the private sector—and then further narrowed down by specific job categories and disciplines.

Academia. The academic employment sector considered here comprises 2- and 4-year colleges, universities, and university-affiliated research institutes. STEM graduates at the bachelor’s level are typically employed as research assistants, research associates, or technicians. Master’s-level graduates are employed predominantly as research associates and staff scientists or, at teaching institutions, as instructors or lecturers. The minimum requirement for a tenure-track professor position is a Ph.D., with many positions now even requiring one or more postdoctoral appointments (postdocs). We found no literature proclaiming a shortage of STEM graduates in the academic employment sector. On the contrary, numerous articles bemoan the lack of permanent faculty positions—a state of affairs that forces young Ph.D.’s to take low-paying temporary positions as postdocs and adjunct faculty.

Many students enter doctoral programs with the intent of climbing the academic ladder and obtaining tenure as a professor. But in many fields, open positions are difficult to find. In fact, the intensified competition for assistant professor openings has resulted in higher quality new hires, meaning that there is a greatly increased chance of obtaining tenure. Then, as the probability of achieving tenure increases, the number of new slots will decline, further exacerbating the shortage of STEM faculty slots.

To examine the production of Ph.D.’s for the academic job market, we and a colleague borrowed the concept of R0, the basic reproduction number, and applied it to academia.14 For academia, R0 was defined as the mean number of new Ph.D.’s a typical tenure-track faculty member will graduate during his or her academic career. When R0 = 1.0, each professor, on average, graduates one new Ph.D. that can replace him or her. When R0 > 1.0, the number of faculty slots has remained almost constant and there are more workers with doctorates than there are faculty positions.

Using this method, we estimate R0 for all fields of study in the United States. (We assume the average career duration to be 20 years.15 We use 2012–2013 data from the College and University Professional Association for Human Resources (CUPA-HR), which reports the number of tenured and tenure-track faculty at 794 institutions in the United States.16 We also use data from the National Center for Education Statistics’ Integrated Postsecondary Education Data System, which has the number of Ph.D.’s awarded in 2012 at those same institutions. We group disciplines by their Classification of Instructional Programs (CIP) code,17 a taxonomic scheme devised by the National Center for Education Statistics to track fields of study. Figure 1 shows that R0 varies considerably across the broad disciplines listed.

Although the number of Ph.D.’s has been climbing steadily, the number of professor positions has remained almost constant in most fields, except for the biomedical sciences and computer sciences.18 A higher R0 indicates that more Ph.D.’s are competing for tenured and tenure-track faculty slots, provided that the number of positions remains constant. For example, R0 = 6.9 signifies that a tenure-track position is available for only 14 percent (1 out of 6.9) of new Ph.D.’s in engineering. Our calculations show that R0 > 1 for all STEM fields, indicating that there are more Ph.D.’s eligible for academic positions than there are openings, assuming no growth in the number of tenure-track faculty slots.

These R0 statistics confirm anecdotal accounts. Faculty openings today often attract hundreds of qualified applicants.19 Henry Sauermann and Michael Roach studied the preferences of science Ph.D. students (n = 4,109) and found that the majority considered a faculty research career to be an “extremely attractive” career path.20 However, only a fortunate few go directly from graduate school to a tenure-track faculty position. In 2010, less than 15 percent of new Ph.D.’s in science, engineering, and health-related fields found tenure-track positions within 3 years after graduation.21 For Ph.D.’s in the life sciences, the figure was an even smaller 7.6 percent. Most who want an academic career join academia as postdocs or adjunct faculty, hoping to vie for a tenure-track faculty position in the future.

Our findings here are consistent with many others in the literature. In 2007, Michael S. Teitelbaum highlighted the poor prospects for recent doctorates and postdocs.22 Similarly, the RAND Corporation pointed out that the length of postbaccalaureate study for the biosciences has increased considerably, from between 7 and 8 years to between 9 and 12, and that many are unable to secure stable employment with tenure until their late thirties.23 This finding was substantiated in a National Research Council report, Bridges to Independence, which focused on the poor state of biomedical research careers and urged immediate reform to enhance the quality of training and to foster opportunities for young researchers to conduct independent research.24 Although this academic surplus began in the biosciences, it has now extended to encompass many STEM fields, such as astronomy, meteorology, and high-energy physics.25

Thus, in the academic employment sector, we find no evidence of any shortages. To the contrary, it appears that the mismatch is between an oversupply of Ph.D.’s desiring an academic career and the relative paucity of tenure-track faculty positions.26 Although the degree of mismatch varies according to discipline, we have long queues of Ph.D.’s competing for nearly all STEM-related faculty positions.

The government and government-related sector. For the purposes considered here, this sector comprises different branches of civilian government organizations that require their employees to hold U.S. citizenship and certain security clearances. Examples are the U.S. Department of Energy’s National Laboratories and the U.S. Department of Defense (DOD), the military, and a number of defense and aerospace contractors and research institutes.This section synthesizes reports produced by the National Academies that studied the hiring needs of the U.S. Air Force and the DOD with anecdotal accounts from the authors’ interviews.

The National Research Council Committee states that the Air Force had a robust supply of personnel with STEM degrees to meet its recruiting goals for STEM positions, with a few exceptions.27 The Air Force Personnel Center found staffing gaps in electrical engineering, operations research, quantitative psychology, physics, nuclear engineering, and systems engineering, specifically with regard to graduates with advanced degrees. The Aeronautical Systems Center commander also identified shortages, in areas such as electromagnetics, structures, software, reliability and maintainability, and manufacturing engineering.

Similarly, the National Academy of Sciences Committee, charged with identifying the needs of the U.S. DOD and the U.S. defense industrial base, found that DOD representatives almost unanimously stated that there was no STEM workforce crisis, but that there were specific areas in which needs were not being met.28 For example, 800 funded positions were open for 90 days or more for systems engineers and other STEM workers, and there were opportunities for cybersecurity and intelligence professionals as well. In addition, the aerospace and defense industry has experienced difficulty in hiring mechanical engineers, systems engineers, and aerospace engineers.

These sentiments were generally echoed in our interviews. One participant, a recruiting manager for a government research institute, said that hiring at the bachelor’s level was relatively easy, but hiring those with advanced degrees was proving more challenging because of skill set mismatches.29 He stated that, although there were many applicants from the mechanical, aeronautical, and bioengineering disciplines, shortages of electrical engineers existed at the doctoral level. Software development skills at all degree levels were also in high demand.

Another recruiting manager for a government research institute found difficulties hiring those with advanced degrees in computer sciences and computer engineering.30 Because of budget stipulations, salaries his institute offered could not compete with those in the private sector.

Although foreign nationals can generally be brought in to bridge skill gaps in academia and the private sector, that is currently not an option in many areas for government workers and contractors, including defense-related contractors. The International Traffic in Arms Regulations dictate that information and material related to defense and military technologies may be shared only with U.S. citizens unless a specific exemption is obtained. A manager for a large government contractor found substantial shortages in hiring of Ph.D.’s in fields such as nuclear engineering, materials science, and thermohydraulic engineering.31 This contractor requires only a dozen or so workers in each field, but the supply of U.S. citizens with doctorates in these fields is small.

A recruiter seeking people to work in engineering startup companies told us of problems finding materials science Ph.D.’s who were U.S. citizens.32 Although the recruiter received dozens of applications from qualified foreign nationals, the government funding involved required U.S. citizenship.

In the government and government-related employment sector, we found no evidence of widespread STEM shortages; however, there may be shortages at the advanced-degree level due to citizenship and security clearance requirements.

Private sector. Much of the literature on the STEM crisis emanates from concerns about shortages or surpluses in the private-sector STEM labor market; however, the crisis is generally discussed in broad terms, referencing the STEM workforce as a whole. For example, the report by the President’s Council of Advisors on Science and Technology called for an additional 1 million STEM degrees over the next decade.33 Similarly, many studies dispute the claim that there are STEM shortages at the aggregate level and point to shortages only in specific fields.34 However, the disciplines involved and the degree levels at which graduates are actually in demand are unclear.

The findings that follow are from a literature review and interviews.

1. Shortages. There are many accounts based on anecdotal evidence that break down disciplines to a relatively detailed level and identify specific areas with a shortage of STEM talent. For example, Lou Frenzel identified shortages among analog/linear and radiofrequency/microwave design engineers and skilled programmers.35 Similarly, Jonathan Rothwell analyzed the Conference Board’s Help-Wanted Online Series® and found that in 2010 there were seven job openings in computer occupations for every graduate with a relevant computer major.36 And Abby Lombardi, in Wanted Analytics, which aggregates job listings from all over the World Wide Web, reported in 2013 that help-wanted ads for software developers were up 120 percent over the previous year.37

From our interviews with recruiters, we also find software development skills to be the most in demand. Experienced mobile application developers are especially coveted.38 In certain cases, it does not even matter whether a candidate has a bachelor’s degree in a specific area: companies39 are looking for candidates with hands-on experience in software development through “hack-a-thons,” extracurricular projects, and internships. These anecdotal accounts are supported by a falling unemployment rate for software developers, from 4 percent in 2011 to 2.8 percent in 2012 and down to 2.2 percent in the first quarter of 2013.40 Also, the recent “big data” trend has sparked demand for data scientists in all areas, from health care to retail.41

Because energy prices surged in the last decade and new technologies for the domestic extraction of oil and gas emerged, petroleum engineers are now in high demand, even though that occupation was an unattractive and declining one throughout the 1980s and 1990s.42 As an indicator, the real wages of petroleum engineers have increased.43

Demand for STEM skills also exists below the bachelor’s level. A 2011 survey of manufacturers found that as many as 600,000 jobs remain unfilled because there is a lack of qualified candidates for technical positions requiring STEM skills—primarily production positions (e.g., machinists, operators, craftworkers, distributors, and technicians).44 Some are concerned that very few people are pursuing employment in the skilled trades.45

2. Surpluses. At the same time that shortages exist, there are areas with surpluses of STEM talent—most notably, biomedical Ph.D.’s. An NIH blue-ribbon panel found an increasing number of biomedical Ph.D.’s working in science-related occupations that do not involve research and even that do not require graduate training in science.46 Chemistry and biomedical graduates also have taken a hard hit, due to the downsizing and offshoring of biotechnology, chemical, and pharmaceutical jobs.47 Since 2000, U.S. pharmaceutical companies have cut 300,000 jobs.48 By 2012, downsizing had increased the unemployment rate among chemists to 4.6 percent, the highest in 40 years. One recruiter we interviewed said he found that many chemical engineering college graduates were seeking employment in software development.49 Among young Ph.D.’s, the situation was even worse: just 38 percent of newly minted chemistry Ph.D.’s were employed in full-time, nonpostdoc positions in 2011, down from 51 percent in 2008.50 New chemical engineering Ph.D.’s fared better, with a full-time, nonpostdoc employment rate of 61 percent.

In 2010 and 2011, the unemployment rate for electrical engineers held at 3.4 percent, but it spiked to 6.5 percent in the first quarter of 2013. Although recruiters in the government and government contractor sector had concerns about hiring electrical engineers, these concerns did not surface in our interviews with private sector recruiters, suggesting that the hiring challenge in the government sector is probably due to the U.S. citizenship requirement.

Because the unemployment rate for STEM Ph.D.’s is generally low, a more useful indicator of job market strength is the number of STEM Ph.D.’s who accept potentially permanent positions, compared with those who accept postdocs. A considerable number of physics Ph.D.’s are unemployed, accepting postdocs and other temporary positions (69 percent in 2010, as opposed to 51 percent before the dot-com bust), indicating that the demand for physics Ph.D.’s is not high. (See figure 2.)

3. Geographic differences. There are also regional differences in the labor markets for STEM workers. For example, software developers are in much higher demand in California, Washington State, and New York, a fact that is reflected in their higher wages in those states. (See figure 3.) This trend is seen across different STEM occupations, and areas of demand vary. Petroleum engineers, for instance, are clustered in Texas and Oklahoma. A recruiter for a company in Connecticut stated that one of the primary challenges he faced in hiring software developers was the location of the office, because many qualified candidates were reluctant to relocate to Connecticut.51 Another recruiter mentioned that his company relocated to the Boston area specifically to gain access to the local talent pool, a move that improved recruitment.52


Across all the different disciplines, yes, there is a STEM crisis, and no, there is no STEM crisis. It depends on how and where you look.

For most Ph.D.’s, the United States has a surplus of workers, especially in tenure-track positions in academia. The exceptions are certain fields within industry, such as petroleum engineering, process engineering, and computer engineering, and other fields in the government sector, such as nuclear engineering, materials science, and thermohydraulic engineering. Academia tends to absorb the Ph.D.’s who are unable to find positions in industry into postdoc positions. At the bachelor’s and master’s levels, there is consistent demand for employees in software development, as well as in high-growth areas such as mobile application development, data science, and petroleum engineering. There is also demand below the bachelor’s level in the manufacturing industry, which needs workers in the skilled trades, such as machinists and technicians. Hence, we have a heterogeneous mixture of supply and demand for different occupations: some have a queue of workers, others a queue of unfilled positions.

Our findings are supported by the National Center for Education Statistics’ longitudinal study of baccalaureate holders, a survey which found that 69.7% of graduates who had not enrolled in advanced-degree studies after they completed their bachelor’s degrees in the 2007–2008 academic year were employed in a full-time job with an annualized median salary of $46,000 between graduation and 2012.53 For STEM majors, the full-time employment rate increased to 77.2 percent and the median salary was $60,000. However, not all STEM majors were equally in demand: computer and information sciences majors and engineering and engineering technology majors had full-time employment rates of 77.1 percent and 83.2 percent, respectively, and corresponding median salaries of $66,000 and $67,000, while graduates who majored in the biological and physical sciences, science technology, mathematics, or agricultural sciences had a full-time employment rate of 71.4% with a median salary of $46,800, closer to that of non-STEM majors. These data are consistent with our conclusion that there is significant variation in the demand for graduates, depending on the STEM discipline.


This article draws upon a variety of data sources—professional science and engineering societies, labor market data, the National Science Foundation, literature reviews, and anecdotal accounts—to understand the supply-and-demand landscape for the STEM labor market. The analysis presented offers a first cut at identifying disciplines and degree levels that are either in demand or oversupplied. A clearer picture of the supply and demand of the STEM workforce will require better data and consistent monitoring of both employer requirements and STEM worker availability.

We introduced the taxicab queuing model as a metaphor for the STEM labor market. Depending on the STEM job segment, we can either have a queue of positions waiting to be filled (cf. taxis) or a queue of STEM workers waiting for jobs (cf. passengers). The characteristics of the queue depend on different factors: the rate of job turnover (cf. taxi service rate); the STEM worker arrival rate (cf. passenger arrival rate); the number of positions available (cf. the number of taxis in the fleet); the location of the job; the degree held by the worker (cf. type of taxi); and the worker’s citizenship status. The model also highlights the probabilistic nature of the supply-and-demand market: random fluctuations can cause job segments that traditionally have a shortage of workers to have a surplus, and vice versa. Although we currently lack the data to operationalize the model, it presents a novel approach to characterizing the variation across STEM job segments.

Our central question is whether there is a “STEM crisis” or a “STEM surplus.” The answer is that both exist. Our analysis yields the following findings:

  • The STEM labor market is heterogeneous. There are both shortages and surpluses of STEM workers, depending on the particular job market segment.
  • In the academic job market, there is no noticeable shortage in any discipline. In fact, there are signs of an oversupply of Ph.D.’s vying for tenure-track faculty positions in many disciplines (e.g., biomedical sciences, physical sciences).
  • In the government and government-related job sector, certain STEM disciplines have a shortage of positions at the Ph.D. level (e.g., materials science engineering, nuclear engineering) and in general (e.g., systems engineers, cybersecurity, and intelligence professionals) due to the U.S. citizenship requirement. In contrast, an oversupply of biomedical engineers is seen at the Ph.D. level, and there are transient shortages of electrical engineers and mechanical engineers at advanced-degree levels.
  • In the private sector, software developers, petroleum engineers, data scientists, and those in skilled trades are in high demand; there is an abundant supply of biomedical, chemistry, and physics Ph.D.’s; and transient shortages and surpluses of electrical engineers occur from time to time.
  • The geographic location of the position affects hiring ease or difficulty.

As our society relies further on technology for economic development and prosperity, the vitality of the STEM workforce will continue to be a cause for concern.

ACKNOWLEDGMENTS: Work on this project was supported by the National Institutes of Health (NIH) under a grant titled “Developing a Scientific Workforce Analysis and Modeling Framework (SWAM),” awarded to The Ohio State University and the Massachusetts Institute of Technology (MIT), Center for Engineering Systems Fundamentals (CESF)—NIH Grant # 5 U01 GM094141-02. The discussion and conclusions presented are those of the authors and do not necessarily represent the views of the NIH, The Ohio State University, or MIT. We thank Joshua Hawley of The Ohio State University and Navid Ghaffarzadegan of Virginia Tech for helpful comments on an earlier draft.


1 President’s Council of Advisors on Science and Technology, Engage to excel: producing one million additional college graduates with degrees in science, technology, engineering, and mathematics (Executive Office of the President of the United States, 2012).

2 Michael Anft, “The STEM crisis: reality or myth?” The Chronicle of Higher Education (November 11, 2013).

3 B. Lindsay Lowell and Harold Salzman, Into the eye of the storm: assessing the evidence on science and engineering education, quality, and workforce demand (Washington, DC: Urban Institute, October 29, 2007); Anthony P. Carnevale, Nicole Smith, and Michelle Melton, STEM: science, technology, engineering, mathematics (Washington, DC: Georgetown University Center on Education and the Workforce, 2011); and Terrence K. Kelly, William P. Butz, Stephen Carroll, David M. Adamson, and Gabrielle Bloom, eds., The U.S. scientific and technical workforce (Santa Monica, CA, Arlington, VA, and Pittsburgh, PA: RAND Corporation, June 2004).

4 The taxicab queue is a classic queuing theory problem that models the queues for taxis and passengers as a function of the arrival rates of passengers and taxis and the size of the taxi fleet. The arrival rate of passengers is modeled as a Poisson process, and the arrival time for a taxi is modeled as a conditional Poisson process which depends on the number of taxis that are currently busy. Numerical and graphical results of the taxicab queuing model can be found in Yi Xue, “STEM Crisis or STEM Surplus?” master’s thesis, Technology and Policy Program (Cambridge, MA: Massachusetts Institute of Technology, 2014).

5 Tapping America’s potential: the Education for Innovation Initiative (Washington, DC, Business Roundtable, 2005); Ensuring a strong U.S. scientific, technical, and engineering workforce in the 21st century (Washington, DC: National Science and Technology Council, April 2000); The science and engineering workforce: realizing America’s potential (Arlington, VA: National Science Board, 2003).

6 National Academy of Sciences, National Academy of Engineering, and Institute of Medicine, Rising above the gathering storm: energizing and employing America for a brighter economic future (Washington, DC: The National Academies Press, 2007), pp. 1–591.

7 Senator Bob Casey, STEM education: preparing for the jobs of the future (Washington, DC: U.S. Congress Joint Economic Committee, April 2012).

8 Engage to excel.

9 “Testimony of Michael S. Teitelbaum before the Subcommittee on Technology and Innovation,” November 6, 2007; see transcript,

10 Lowell and Salzman, Into the eye of the storm.

11 Hal Salzman, Daniel Kuehn, and B. Lindsay Lowell, Guestworkers in the high-skill U.S. labor market: an analysis of supply, employment and wage trends, EPI Briefing Paper no. 359 (Washington, DC: Economic Policy Institute, April 24, 2013), pp. 1–35.

12 Newsroom: Census Bureau reports majority of STEM college graduates do not work in STEM occupations (U.S. Census Bureau, 2014). The U.S. Census Bureau includes sales and managerial occupations, as well as social science occupations, in its definition of STEM occupations.

13 David George Kendall, “Some problems in the theory of queues,” Journal of the Royal Statistical Society, Series B, vol. 13, no. 2, 1951, pp. 151–185,

14 Richard C. Larson, Navid Ghaffarzadegan, and Yi Xue, “Too many PhD graduates or too few academic job openings: the basic reproductive number R0 in academia,” Systems Research and Behavioral Science, November/December 2014, pp. 745–750.

15 See Richard C. Larson and Mauricio Gómez Díaz, “Nonfixed retirement age for university professors: modeling its effects on new faculty hires,” Service Science, March 2012, pp. 69–78.

16 Although only 294 of the 794 institutions have doctoral programs, the number of faculty used in the calculation of R0 is the total for all of those institutions, because the positions referred to are still tenured and tenure-track faculty positions.

17 Disciplines were included only if there were data available for both the number of Ph.D.’s and the number of faculty.

18 National Science Board, Science and engineering indicators 2014 (Arlington, VA: National Science Foundation, 2014).

19 Beryl Lieff Benderley, “The real science gap,” Pacific Standard, June 14, 2010,

20 Henry Sauermann and Michael Roach, “Science PhD career preferences: levels, changes, and advisor encouragement,” PLOS One, May 2, 2012,

21 National Science Board, Science and engineering indicators 2014.

22 Michael S. Teitelbaum, The current model of STEM graduate education and postdocs: is it evolving to meet needs of the nation and its participants? paper presented at the National Institutes of Health, Washington, DC, November 5–6, 2007,

23 Kelly et al., eds., The U.S. scientific and technical workforce.

24 Bridges to independence: fostering the independence of new investigators in biomedical research (Washington, DC: National Research Council, 2005).

25 Benderley, “The real science gap.”

26 Navid Ghaffarzadegan, Joshua Hawley, Richard C. Larson, and Yi Xue, “A note on PhD population growth in biomedical sciences,” Systems Research and Behavioral Science, November 21, 2014.

27 National Research Council, Examination of the U.S. Air Force’s science, technology, engineering, and mathematics (STEM) workforce needs in the future and its strategy to meet those needs (Washington, DC: National Academies Press, 2010).

28 National Academy of Sciences, Report of a workshop on science, technology, engineering, and mathematics (STEM) workforce needs for the U.S. Department of Defense and the U.S. defense industrial base (Washington, DC: National Academies Press, 2012).

29 Research institute A, involved primarily in U.S. government projects that require U.S. citizenship. (Here and hereafter, sensitive organizations are identified by a letter abbreviation.)

30 Research institute B, involved primarily in U.S. government projects that require U.S. citizenship.

31 Engineering company A, a government contractor that requires U.S. citizenship and hires no dual-citizenship holders.

32 Engineering startup.

33 President’s Council of Advisors on Science and Technology, Engage to excel.

34 For example, Lowell and Salzman, Into the eye of the storm; see also “Testimony of Michael S. Teitelbaum.”

35 Lou Frenzel, “Is there really a shortage of engineers?” Communiqué, September 16, 2013,

36 Jonathan Rothwell, “The need for more STEM workers,” The avenue: rethinking metropolitan America (Washington, DC: The Brookings Institution, June 1, 2012),

37 Abby Lombardi, “Software development ranks as the most in-demand skill for tech jobs,” Wanted Analytics, September 5, 2013,

38 Media company A.

39 Media company A, engineering company B, and information technology company A.

40 Patrick Thibodeau, “Electrical engineers see sharp uptick in Q1 jobless rate,” Computerworld, April 22, 2013,

41 Catherine de Lange, “So you want to be a data scientist?” Naturejobs, March 18, 2013,

42 See John Porretto, “Oil companies scramble to find engineers,” USA Today, August 31, 2007,; Vivek Wadhwa, “Mr. President, there is no engineer shortage,” The Washington Post, September 1, 2011; and Michael S. Teitelbaum, “The myth of the science and engineering shortage,” The Atlantic, March 19, 2014.

43 Salzman, Kuehn, and Lowell, Guestworkers in the high-skill U.S. labor market.

44 Tom Morrison, Bob Maciejewski, Craig Giffi, Emily Stover DeRocco, Jennifer McNelly, and Gardner Carrick, Boiling point? The skills gap in U.S. manufacturing (Washington, DC, and New York: The Manufacturing Institute and Deloitte, 2011).

45 See Joshua Wright, “America’s skilled trades dilemma: shortages loom as most-in-demand group of workers ages,” Forbes, March 7, 2013.

46 Biomedical research workforce working group report (Department of Health and Human Services, National Institutes of Health, 2012),

47 See David Cyranoski, Natasha Gilbert, Heidi Ledford, Anjali Nayar, and Mohammed Yahia, “The PhD factory,” Nature, April 20, 2011, pp. 276–279; and Jonathan Bloom, “America’s vanishing science jobs,” New York Post, June 24, 2011.

48 Brian Vastag, “U.S. pushes for more scientists, but the jobs aren’t there,” The Washington Post, July 7, 2012.

49 Information technology company B.

50 Susan R. Morrissey, “Starting salaries,” Chemical and Engineering News, June 4, 2012, pp. 36–38.

51 Media company A.

52 Information technology company C.

53 Emily Forrest Cataldi, Peter Siegel, Bryan Shepherd, Jennifer Cooney, and Ted Socha, Baccalaureate and beyond: a first look at the employment experiences and lives of college graduates, 4 years on (B&B:08/12), NCES 2014-141 (National Center for Education Statistics, Institute of Education Sciences, July 2014).

About the Author

Yi Xue

Yi Xue is a former graduate student in the Technology and Policy Program, Massachusetts Institute of Technology, Cambridge, Massachusetts.

Richard C. Larson

Richard C. Larson is Mitsui Professor, Engineering Systems Division, and founding director, Center for Engineering System Fundamentals, Massachusetts Institute of Technology, Cambridge, Massachusetts.

close or Esc Key