The EPI study keeps misleading people. Key points to remember:
(1) The EPI is a unionized-labor think tank with strong anti-immigration views. Its research – its selection of figures and spin – will always be in service of that goal.
(2) This latest EPI report, like a similar one before it, uses industry-wide and degree-wide averages with no adjustment for cohorts of experience/age. It is thus possible for the 'average' to go down if there is an in-rush of new younger workers (or equivalently, less-experienced career-changers at starting salaries).
As a simple contrived example, assume an 'industry' starting with two workers, a entry-level worker at $50K a year, a senior worker at $100K a year. The average is $75K.
A year passes. Business is booming and there's a shortage of well-qualified workers. The senior worker gets a raise to $125K. The now-junior worker gets a raise to $62.5K. The company would like to hire more senior stars, but they're all already doing well in other industries (or other countries and deterred from working domestically by immigration barriers). Still, the sector is growing so they hire two new promising entry-level workers at $50K each, one recent graduate and one career-switcher, who were both making just $40K last year in non-STEM work. Maybe one of them will be a domain-specific star, with salary to match, once trained-up in a year or so.
Every single person in this scenario got a 25% year-over-year raise... but the 'industry average' compensation dropped from $75K to $71.9K, split over a larger and earlier-in-career base.
If you live in the simpleminded, interchangeable-STEM-laborer world of the EPI analysis, this is a crisis. There are plenty of "STEM workers" because every "STEM worker" is just like every other "STEM worker" and the new entrants haven't magically started making the old $75K average simply by putting on a STEM-worker-jumpsuit. But really, everyone's doing better, the industry is talent-constrained and responding with both salary increases and legally-allowed hiring, and would be able to expand more if other truly 'senior' industry-skilled people were available.
(3) The EPI report uses wordcrafting like "[f]or every two students that U.S. colleges graduate with STEM degrees, only one is hired into a STEM job" to create the impression many STEM graduates are left without a desired job. But there's plenty of evidence (including in their own report) that STEM-graduate unemployment remains low, and lots of workers happily go into non-STEM jobs with STEM degrees, or into STEM jobs with non-STEM degrees. (In particular, many STEM graduates prefer the pay and challenges of other fields – law, finance, consulting, even many sales and small-business roles – and may have taken their STEM degree simply out of intellectual interest, not as strict occupational training.)
So the healthy churn of matching of people with jobs where they, individually, thrive gets spun by EPI as implying we have more STEM graduates as we need. They seem to think of a STEM degree as if it were chiefly a 'union card' giving you the right to work in some rationed job role... and if not everyone with these cards takes those jobs, too many cards were issued. Cut back on the STEM worker supply, from schools and immigration, pronto!
That's the kind of fixed-pie, world-oblivious analysis that did wonders for the big-three US auto industry... don't let the EPI port that thinking into the high-tech world.
We need as many smart, trained workers - from schools and career-changers, foreign and domestic – as we can get.