Thanks for posting this. But that fits Nate's point- you look for jumps, not sustained rates, and that sustained high rates do not equate to losing. 72-76. Loss. 88-92. Loss. But 80-84? Win. And the point was that there's no correlation to winning or losing by having high numbers. You just can't have the number spike.
The 76-80 number doesn't fit what Silver posted, so I'd like to know what he was using there as opposed to what you've posted. But the lack of causation is still readily apparent.
Edited to add: He sidestepped it, which is intellectually dishonest (especially since a single event doesn't disprove the point--that high unemployment doesn't mean you lose):
fivethirtyeight Nate Silver
Over the past 100 years, no incumbent president with an unemployment rate of between 7.6% and 23.5% has failed to win re-election.
Still doesn't explain Ford. Unemployment was actually down from 75 to 76. Perhaps using the term "re-election" is more sophistry?
It still looks to me like at best those figures take in no one post WWII, in which case they are useless.
But let's look at the jump theory.
Nov 1948 -- 3.8%
Nov 1952 -- 2.8% -- -1% -- Incumbent did not run for re-election, highly unlikely he would have won though; opposition party won.
Nov 1956 -- 4.4% -- +1.6% -- Incumbent won
Nov 1960 -- 6.2% -- +1.8% -- no incumbent, incumbent party lost.
Nov 1964 -- 4.8% -- -1.4% -- appointed incumbent won
Nov 1968 -- 3.4% -- -1.4% -- incumbent did not run for re-election, highly unlikely he would have won; opposition party won
Nov 1972 -- 5.3% -- +1.9% -- incumbent won
Nov 1976 -- 7.8% -- +2.5% -- appointed incumbent lost
Nov 1980 -- 7.5% -- -0.3% -- incumbent lost
Nov 1984 -- 7.2% -- -0.3% -- incumbent won
Nov 1988 -- 5.3% -- -1.9% -- no incumbent, incumbent party won
Nov 1992 -- 7.4% -- +2.1% -- incumbent lost
Nov 1996 -- 5.4% -- -2.0% -- incumbent won
Nov 2000 -- 3.9% -- -1.5% -- no incumbent, incumbent party lost
Nov 2004 -- 5.4% -- +1.5% -- incumbent won
Nov 2008 -- 6.9% -- +1.5% -- no incumbent, incumbent party lost
May 2011 -- 9.1% -- +2.2%
Incumbent wins -- +1.6, -1.4, +1.9, -0.3, -2.0, +1.5
Incumbent losses -- +2.5, -0.3, +2.1
No incumbent, incumbent party loses -- +1.8, -1.5, +1.5
No incumbent, incumbent party wins -- -1.9
Eligible incumbent does not run, incumbent party loses -- -1, -1.4
About the only thing you can say is that no incumbent has won with more than a 1.9% jump in unemployment from their initial election month to election day. They're 0 for 2 in that scenario, and that's where Obama is right now.
The other thing you might say is that both of GWB's elections might be the most unlikely results in this set. Winning opposing the incumbent party coming off -1.5% unemployment, then winning off +1.5% unemployment without really a major accomplishment (the 1.6% jump was still only to 4.4, and Nixon's re-election came with winding down Vietnam, and establishing relations with China).
But I really don't see a ton to go off here, regardless. Silver's statement is still useless, to say the best about it.