Fair enough. I should have been more specific. My intent is that it is people of working age. This needs to also account for status of school and retirement. However, we should be able to get trivial numbers for these from SS and college reporting to get an fairly accurate number.
A lot of comments here are heavily biased towards accurately accounting for the retirement end of life, without considering the other end - people who are intentionally not working or are under-employed for some other reason. The one I had mainly in mind was highschool and college-aged students who are working part-time: they'd get included in under-employment, but aren't necessarily looking for a living wage yet because they have other priorities, so they don't get included in unemployment counts.
Also,
> I'd like to see unemployment measured more like total population minus those employed.
This measure would include children who can't even work part-time.