If you go to the Bureau of Labor Statistics data, it is really easy to demonstrate that the duration of unemployment is directly related to industry. That's no big whoop, because we all know that there are cycles in every business. Take, for instance, the mortgage universe. Five years ago there weren't enough workers on the face of the earth to satisfy the industry's appetites for workers. Today, they're a dime a dozen. Another good example drills down one layer deeper...from industry to actual skill set; how about being a cobalt programmer pre 12/31/99 vs. post 12/31/99. Boom to bust. But often they're not that obvious. You need to drill sideways. Look at categories of jobs, and you will always find certain job descriptions that are like the kiss of death (in terms of how long it takes) when it comes time to finding a new job. An example would be a human resources middle management type. There are a lot of them out there in a market where there are fewer jobs available that match their skill set. Whoaa... "human resources" and "skill set" in the same sentence. Now that's a bit oxymoronic!! Sorry, my cynicism runneth over. I'm not disguising what I've been through very well, am I?
My point here is that there seems to be significant data available to an underwriter to understand the myriad of individual risks inherent in underwriting severance insurance. The challenge is automating its retrieval and layering it on a real workforce/employee base.