Wilma: "If they are currently working, they are disloyal. The unemployed may have been laid off for no fault of their own. Don't interview the employed."
Bjarne: "...Maybe instead of guessing, we should just interview the ones that seem qualified, and if they're unemployed we can ask what happened at their last job, and if they're employed we can ask why they're looking for a new job?"
I'm not saying this is morally right, or even smart; but it explains the observed behavior at least as well as "they would be a problem" (which, as you point out, isn't that strong an argument.)
If we can filter out a large percentage of the unskilled with a litmus test like Fizzbuzz that should be a big win. Bigger, in fact, because we've stipulated that the unskilled are far more common than the skilled in today's labor market. I would think that this would be a much better solution than throwing out all resumes from the unemployed.
They also had a puzzle that you didn't have to code for, but had to tell HOW you could code for it.
I think it took more time than you should ask of an applicant (they sure can't apply very many places!) but it definitely weeds out the unqualified. (And possibly the qualified that won't jump through hoops to work for average wages.)
They often lack the scope of the target domain, but that does not equate to a lack of common sense.
they were largely of above average intelligence and aptitude