ALERT - 🚨 – tweetstorm below about my new paper “STEM Careers and Technological Change”, with @knoray23 - link here nber.org/papers/w25065; 1/N
The STEM sector is critical for innovation and economic growth. And everyone knows about the high initial payoff to majoring in STEM 2/N
Yet despite this high payoff, the perception of a “STEM shortage” is widespread. And there are lots of programs and policies designed to get more folks into STEM careers; 3/N
On the other hand, more than half of STEM majors end up in non-STEM jobs. So is there really a shortage? Yes - of STEM skills, not workers per se. 4/N
STEM jobs change rapidly, and new job tasks always need to be learned. That’s hard for older workers. STEM majors learn job-relevant skills that have an immediate payoff. But when the job changes, old skills become obsolete; 5/N
The main finding of the paper is that the return to an applied STEM major starts high but falls by more than 50% in the first decade of working life 6/N 
Declining returns hold for applied STEM majors (comp sci and engineering) but not for “pure” STEM majors like bio, chem, physics and math; 7/N
The declining return to an applied STEM major coincides with a rapid exit from STEM jobs over the age 23-35 period; 8/N
We look at major-by-job interactions – bottom line is flatter wage growth is a feature of STEM jobs, not STEM majors; 9/N
This pattern holds in multiple data sources and when controlling for aptitude and other important determinants of earnings; 10/N
We directly measure job change using online job vacancy data collected by @Burning_Glass. This allows us to construct a detailed picture of job skill requirements and how they change over the 2007-2017 period; 11/N
STEM jobs indeed have higher rates of change than other occupations, and this is driven by the rise and fall of specific software skills and business processes required by employers; 12/N
Consistent with our story, jobs with higher rates of skill-change – both STEM and non-STEM – have flatter age-earnings profiles and employ younger workforces; 13/N
We also find that higher-aptitude workers choose STEM careers initially, but are more likely to exit them over time. This is because the return to aptitude over a career is higher in jobs that don’t change, where knowledge can accumulate 14/N
Why should you care? Three reasons: 1) New facts about STEM careers that you may not have known, and a framework for understanding them; 15/N
2) We show how “big data” can be used to understand technological change at the job level (and in principle, we could dig further into industries, local labor markets, firms…) 16/N
3) STEM degrees are high-skilled vocational education. They definitely pay off, but not as much as short-run comparisons suggest. Non-STEM majors catch up pretty quickly. Don’t undersell the benefits of a liberal arts education! 17/N
Most importantly, the impact of rapid technological change depends critically on our ability to train the next generation of STEM workers and to support “lifelong learning” among the currently employed. 18/N
Here is a link to the NBER working paper. nber.org/papers/w25065
Here is an ungated version. scholar.harvard.edu/files/ddeming/…
Thanks for reading! END 19/19
Here is an ungated version. scholar.harvard.edu/files/ddeming/…
Thanks for reading! END 19/19
