Job applications are a perfect example of an adverse selection problem. Bad candidates are more likely to laid off, so they enter the job market more often. Bad candidates stay on the market longer than good ones, because they have trouble getting hired. Because they're on the market longer, bad candidates apply for more jobs than good ones.
These ugly truths guarantee that 99% of your applications are from bad candidates: people you would never, ever employ, under any circumstance. The sooner you can sort out the remaining 1%, the better. It's expensive to evaluate candidates, and each stage of evaluation is more expensive than the last. (keyword scan, human reading a resume, HR screen, phone screen, interview...)
If you use crude, ugly heuristics that flag false positives on half of your target population, but you save labor by not evaluating 85% or 90% of the applicants, you come out way ahead. Rational companies are bound to throw out a few babies with millions of gallons of bathwater. Cruel, but unavoidable.