Your basic calculation is still right although I'd add more numbers to the opposite site. "Potential graduates" with or without CS can potentially enter to any software sector, or even hardware. Though this is just a small factor.
Anyway having said that, good studies check for shortages and overproductions in more complicated ways than just totting up numbers blindly.
My point was that the major sources of the "no shortage" or particularly "surplus" claims are doing shoddy math that over-represents the labor pool and under-represents demand. Also, more importantly, they're conflating all types of programming.
There are certainly skillsets/interests where there's no shortage, and there are certainly spaces where there is a shortage. People who don't differentiate 'programming' any further aren't saying something useful, and people who do only disagree on the details.
(I realize good studies check for those things, but that's not what I'm talking about. I'm talking about people saying "but aren't there lots of spare programmers we don't need?" because some hack with an infographic subtracted two numbers and found a surplus. I've seen this error pattern in respectable news on a bunch of occasions.)
It was about 7-12% a couple of years ago for grads from CS, Engineering, and IS, at a time when the national level was lower.
https://www.theatlantic.com/education/archive/2014/03/the-my...
I'm not sure if we're on the same page, but my contention is that western governments are encouraging kids to pursue STEM, knowing that there's already an oversupply of workers due to visas issued to emerging economies, and that the issue is compounding year on year.
I object to some specifics - IS without programming education is quite different from CS/programming - but it's still a potent statistic. More narrowly, my (cynical) take is that CS degrees are decidedly unequal. Having seen multiple companies say they only bother with top-seven job fairs, and some genuinely talented friends struggle to land interviews with third-string universities on their resumes, I get the sense that the "shortage" and "surplus" are capable of existing side-by-side.
I completely agree with your conclusion, since it holds regardless of whether I'm right about segmentation within CS degrees: a lot of people are being unhelpfully pushed into "employable" STEM degrees by people who want to look science-friendly, and they're going to be left holding the (unemployment) bag when that doesn't work out.
> I get the sense that the "shortage" and "surplus" are capable of existing side-by-side.
I guess this is fundamentally the issue
I see a bait and switch conducted here all the time, where thinkpiece writers will count up the number of annual graduates in Computer Science (or sometimes Information Technology in general), subtract the number of new software jobs per year, and conclude that we have massive overproduction of programmers.
This is obviously stupid, but I keep seeing it. It ignores CS graduates who stay in academia or take nonprogramming jobs (e.g. in finance). Often it counts "new software engineering positions", a metric which ranges from insufficient (by excluding Sysadmins and similar) to idiotic (by excluding "programmers" and other job titles the author couldn't be bothered to search for). Shockingly often, it counts new positions but doesn't account for positions vacated by people leaving programming (e.g. to management or retirement). And, of course, it ignores employability, pushing an implication that any CS graduate can fill any software job regardless of specialization, experience, or interest.
It is a relentlessly stupid 'statistic', but I think one reason for this ongoing perception of surplus talent is that people keep rewriting the same silly article.