The SGS numbers are basically fictional. See this contemporaneous discussion of what really changed between the pre-1994 and post-1994 counting methods: http://www.bls.gov/mlr/1995/10/art3full.pdf. The current U4-U6 do include discouraged workers, they simply require the person to indicate on the survey that they have searched for a job in the past year rather than the unworkably vague prior methodology that simply asked whether the person wanted a job. See: http://www.ritholtz.com/blog/2012/08/discouraged-by-the-medi....
The most sensible employment measure is U4 (official unemployment + people who have searched for a job in the last year). U5 and U6 are far too sensitive to demographic changes to serve as a good basis for comparison. E.g. since the 1950's the work force has added a lot of female part-time workers that are counted in U6, but would have been stay-at-home wives in the 1950's and thus not counted as unemployed at all. Also, older people have always been over-represented among the marginally-attached, part-time work force, and as the population ages there are naturally more of this demographic. Both trends inflate U6, but neither are a negative indicator for the economy per se.
Unemployment in the US is near 15% including U6, or near 23% if using pre-1994 counting methods: http://www.shadowstats.com/alternate_data/unemployment-chart...
(On that subject, I'm looking for work in the north SF bay area, Sonoma/Marin region)