Ben Casselman has a fascinating dive
into the long-term unemployment data at the new 538 site. He finds that the long-term unemployed are driven in large part by luck. An unemployed person is more likely to be unemployed for a long period of time when they happen to lose their job at a time of high unemployment. Here’s their core chart:
He also finds that this effect is stronger for those who are unlikely to receive unemployment insurance.
One comment I had. There’s an argument that the long-term unemployed are the weakest employees, those who were fired during the first wave of layoffs that started in 2008. These workers were going to have a hard time finding jobs not based on the labor market but because, to be blunt, they weren’t good workers. (One manifestation: Tyler Cowen did a lot with this idea of zero marginal product workers, ignoring that the marginal product of labor is impacted by demand, back in 2011.) Since long-term unemployed workers look a lot like the general unemployment pool, this is thought to be driven by softer, not-quantifiable, worker characteristics.
If that was the case, then the job losers on the upswing of unemployment, during the first wave of layoffs in 2008 when unemployment was in the 5-8% range, should be more likely to have become a member of the long-term unemployed. They should even be worse than those leaving their job when unemployment was 10% in fall 2009 (which was technically 3 months after the recession ended). But we see a pretty consistent pattern in that chart, which tentatively give evidence that it’s not just the initial skill level of the workers driving the level of long-term unemployment.