If I respond to your post piecewise I will end up babbling on, so I’ll try to address it all in a way in which it is obvious which parts I am addressing. Here goes nothing…
You absolutely do not need a PhD to work in research. There are a lot of cases where it helps, but it is not any kind of universal requirement, even if you work at a “research shop.” In academia, the brand of research that is practiced generally spans from basic to the earliest stages of applied research where the processes are essentially the same and a PhD is basically required.
National labs, on the other hand, are somewhat different. As someone who presently works at one, I have seen that this varies quite substantially depending on your division within the labs. In my division, it is very heavily research focused and almost everyone has a PhD. On the other hand, I have a good friend with only a MS working in high-performance computing who has told me that almost everyone there only has a MS and that getting a PhD is, by and large, completely unnecessary.
In industry, there are often more MS-holders than there are PhDs, as their research generally spans from applied up to the point just before production. Not as much of the work requires the skills a PhD brings to the table.
Regarding the monetary examples, I am aware of the dangers of painting in broad strokes, but every example you have given so far has been in the context of money. I responded in kind. And yes, I have absolutely come across plenty of students have viewed the advanced degree as just a rubber stamp on their way up the corporate ladder. Do I think all of them do this? No, and it isn’t even the majority, but it does happen relatively frequently. There is absolutely an opportunity cost associated with everything, and it is not always monetary. Sometimes it is, but sometimes it is psychological or involves other factors, and assessing those costs on an individual basis does not always result in a thesis-based degree being useless or somehow a bad decision.
Your position seems to rest on the assumption that there are no opportunities available to someone with a thesis-based MS that wouldn’t be better served by either a PhD or a non-thesis MS, and that simply isn’t the case. For some (myself included), they go do an internship and see that those with just a BS are doing jobs that are just absolutely boring to them and that in order to get the jobs they want, they need additional sets of skills not typically available to an undergraduate. My example has been CFD so I will keep using it. You could either just take a job with your BS and start learning CFD while performing your other, boring job duties, or you could decide to go to graduate school and learn the skills you seek there. I don’t think there is a universally correct answer. For me, I would have rather gone back to school instead of crushing my spirits with a job that is soul-crushing to me. Other people may come to a different conclusion. My entire point here has simply been that there is no universally correct answer and that thesis-based degrees can be useful.
Also, I think your points about uncertainty and stability are quite important, but 1 additional year for a thesis-based MS is not likely to meaningfully affect these issues. They are much more prominent among PhD students (I know, I’ve lived that debacle). The caveat to my position here is that I am aware that sometimes thesis-based MS degrees do take longer, and for that reason I advocate that students have a serious discussion with potential MS advisors about the expectations for graduation and the expected time to completion. I think many instances of 4 year MS degrees come down, at least a little bit, to irresponsibility on the part of the advisor and that this should be an important data point in deciding where to pursue the MS. After all, if you are going to spend 3+ years getting an MS, you might as well just stick around for the PhD.
Regarding undergraduate research as a path to these skills, it’s tough to paint with a broad brush here. The quality of undergraduate research (based on my time in graduate school and now at the national lab) seems to be very variable. Factors include the individual students, the projects assigned by PIs, and the graduate mentors the students end up with, but ultimately, a fair bit of that comes down to luck. It’s unusual (though not impossible) that undergraduate researchers reach anywhere near the level of competence that would seem to make them a shoo in for jobs requiring those specific, more advanced skills like CFD.
Finally, I very much take issue with your description of PhD programs and what constitutes learning. What necessary level of competence do you think is actually required for research? Research is learning. The entire concept behind research is adding to the whole of human knowledge, and in order to do that, you have to understand and have a set of learning skills that basically transcends fields. Yes, in the process of earning a PhD you are going to become an expert on some bit of subject matter, but that is really just a result of the above process, not the purpose of the endeavor. The point of a PhD is to prove that you can, if desired, apply this same concept to new areas and generate new knowledge (at least this is the case if your goal is to parlay the PhD into a research position of some kind).
Probably the best example of this is the people I work with here at the national lab. You see people with all manner of technical backgrounds applying their expertise to all manner of new problems, some of which are barely even related to their original field. I have a friend with a background in experimental fluid mechanics who is now also working on a quantum cryptography project. This isn’t because he is a quantum cryptography expert. It’s because he has proven that he is a competent researcher, and research is learning. A PhD is all about learning how to research, and therefore learning how to learn. This is every bit as important to your future career as the subject matter you study with this process.
This kind of misunderstanding also plays into the idea that coursework provides largely the same skills as the thesis or dissertation (or rather the process of the work leading up to writing them). Coursework demonstrates a familiarity with more advanced topics. This is not the same thing as the practical skills that should be gained over the course of working on a thesis. It is especially not the same as the demonstrated ability to lead an original research project from start to finish and contribute to the whole of scientific knowledge that is the point of the PhD. All of these degrees serve their purpose and are useful, but claiming that they all provide essentially the same set of useful abilities is at best disingenuous and at worst illustrates an incomplete grasp of the scope and purpose of each of these degrees.