But what's really so wrong with people dropping out? As you say, the job market is super-saturated, and there just aren't jobs for many. Perhaps it isn't so bad that lots of people do a couple years of a PhD, learn quite a bit, maybe get a MA, and drop out to pursue another career path. I'm going in to a PhD knowing full well that I might not get a job in academia. As a result, I'll be looking quite aggressively at alternative professional opportunities.
I don't think though that requiring a MA, for example, would be all that much of an improvement. Many of these programs are unpaid. Does it really make sense to incentivize more folks to pursue them? There the tradeoff becomes much starker. And for every 1 person who successfully finishes an MA and is accepted to a top graduate program, another 1 still doesn't get in, and is left with just more debt. That's of course exaggerated for effect, but I think there's a serious point there.
Regarding smaller cohorts, I think I largely agree. One worry is that smaller cohorts would encourage universities to cut funding for these departments and for research (which would be terrible). But certainly non-elite elite programs should consider reducing the number of students they take