Let’s say Jack and John are both employed doing the same work while employment is high (and fear of unemployment is therefore low–if either loses his job he can likely find another one). Then the economy craters and Jack loses his job. John’s fear of unemployment is now higher than it was, since he has sees not only his colleague Jack get laid off but also dire unemployment figures in the news. Now fearful, John works harder to keep his job than he did when the economy was good, and by so doing makes up for the productivity lost by Jack’s layoff.
To management, who measures only total productivity and not per-worker productivity, it looks like Jack was a zero marginal productivity worker. As the economy rebounds they’re under no pressure to rehire Jack, so unemployment stays high.
I’d be interested in any reactions to this theory, and in finding out how one would formally describe it in economic terms.