What if these groups don’t overlap?

I.e. the “people who are able to exit” group may not overlap the 5% that would have resigned because they were already searching for something.

The GP’s (codezero) point above still seems valid to me because of that.