October 15, 2018

New Insights on Computer Science and Engineering Careers

New research finds that rapid technological progress in fields like computer science and engineering changes the task content of jobs in those fields substantially over time. That, in turn, causes employers to value recent graduates trained in the newest skills in those disciplines.

David Deming, a Harvard professor of Public Policy, Education, and Economics, presented his research on STEM
careers at an MIT Institute for Work and Employment Research seminar earlier this month.

There is a perception that there is a shortage of qualified workers for science, technology, engineering and math (STEM) jobs in the U.S., David J. Deming, a professor of Public Policy, Education, and Economics at Harvard Kennedy School and Harvard Graduate School of Education, observed during a recent presentation at the MIT Institute for Work and Employment Research (IWER) weekly seminar. “We have this idea that we don’t have enough STEM workers,” he said.

The reality, however, is more complex, as Deming went on to explain. He presented the results of new research he conducted with Kadeem L. Noray, a doctoral student in public policy at Harvard; the research has recently been featured as a National Bureau of Economic Research working paper, “STEM Careers and Technological Change.” As part of their research, Deming and Noray studied the tasks included in online job listings in the U.S. between 2007 and 2017 and looked at the rate at which the task content of occupations changed.

A key finding from that research is that rapid technological progress in applied STEM fields like computer science and engineering changes the task content of jobs in those fields substantially over time. That, in turn, causes employers to value recent graduates trained in the newest skills in those disciplines. As a result, young workers who graduate from college with the newest skills in applied STEM fields are paid a premium over fellow students who had other majors, because the students with applied STEM majors have received very relevant training. “You should think about STEM majors as being high-skilled vocational education,” Deming said.

However, Deming explained, because the task content of jobs in these fields changes quickly, the earnings premium young computer scientists and engineers receive compared to non-STEM majors declines significantly—by more than 50% in the first decade after college graduation. Many STEM workers as a result exit the industry within the first decade after graduation.

In other words, while rapid technological progress creates a perception of a shortage of STEM workers, what’s really in short supply are the newest applied STEM skills. “It is the new job-relevant skills that are scarce, not necessarily the STEM workers themselves,” Deming and Noray write in their working paper. “In fact, faster technological progress contributes directly to the perception of skill shortages by hastening skill obsolescence among older workers.”

Deming and Noray find evidence that people who score better in tests of academic ability are both more likely to enter STEM fields and more likely to leave them in the first decade. Their hypothesis was that young people who are better at learning will initially be drawn to STEM majors and the high-paying jobs they can lead to after college but then will over time be drawn to other fields where the task content of jobs changes less rapidly. Their theory?  In slower-changing fields, workers who learn quickly can accrue knowledge advantages over time—something that’s harder to do in fast-changing fields where knowledge quickly becomes obsolete.

Deming and Noray’s research indicates that different occupations experienced widely differing rates of task change during the 2007-2017 period they studied. Mechanical drafters, the fastest-changing professional occupation as measured by six-digit Standard Occupational Classification codes, saw 40.4% of their tasks change in that decade, while pediatricians, the slowest-changing professional occupation, saw only 3% of their tasks change during that time. After mechanical drafters, the other top ten fastest-changing specific professional occupations were: 2) computer programmers; 3) architectural and civil drafters; 4) software developers, systems software; 5) advertising and promotions managers; 6) environmental engineers; 7) insurance underwriters; 8) pharmacists; 9) electrical and electronics drafters; and 10) actuaries. In contrast, the slowest-changing professional occupations, after pediatricians, were: 2) nurse practitioners; 3) dentists; 4) clinical psychologists; 5) veterinarians; 6) pharmacy technicians; 7) special education teachers, all other; 8) psychiatrists; 9) physicians and surgeons, all other; and 10) nurse anesthetists.)

Deming and Noray’s findings underscore both the importance of lifelong learning during periods of rapid technological change and the “short shelf life” of many technical skills. They conclude:

“Our results inform policy tradeoffs between investment in specific and general educations. The high-skilled vocational preparation provided by STEM degrees paves a smoother transition for college graduates entering the workforce. Yet at the same time, rapid technological change can lead to a short shelf life for technical skills. The rise of coding bootcamps, stackable credentials and other attempts at “lifelong learning” can be seen as a market response to anticipated skills obsolescence. …This tradeoff between technology-specific and general skills is an important consideration for policymakers and colleges seeking to educate the workers of today, while also building the skills of the next generation.”

—Reported by Martha E. Mangelsdorf