The historical trend I

["'Oeps!' De blunders van Rick Perry Indicted"]

As I mentioned in the other thread, the Presidents who ran and lost their reelection all had a smaller margin of victory in their first election than Obama did.

So, how do you explain the case of Bush I, who received 53.37% of the popular vote in 1988, but was defeated in 1992, while Obama received 52.87% of the popular vote in 2008?

Hmm...

Perot.  Factor him out of the race and H.W. at a minimum ekes out a narrow victory in a close contest. 

Bush/Quayle: 273
Clinton/Gore: 265

How in the world does a president with a sub-40% approval rating eke out a win over a relatively strong Democratic ticket? How does he win two states that Dukakis won in '88? Winning CT and ME is also a little far-fetched, considering how hard New England was hit by the recession and the fact that Bush actually finished third in ME. 

["'Oeps!' De blunders van Rick Perry Indicted"]

Check the numbers here - Bush was at 39% in mid-February, before Perot was really a factor. A month later, an ABC/WAPO poll had Bush losing to both Clinton and Tsongas - again before Perot had really caught fire. He had a very mild rebound later in the spring, at the same time Clinton was struggling in the primaries, but that petered out after the LA riots and some bad economic news (the June '92 unemployment rate increased to 7.8%). Perot's success was really a byproduct of Bush's unpopularity (and doubts about Clinton) rather than the driver of it. 

["'Oeps!' De blunders van Rick Perry Indicted"]

I still maintain that if Perot had not entered the race, Bush could have kept those doubts about Clinton uppermost in the public mind.

Ok, but the fact that you have to give him states that Dukakis won with 51% and 55% just to barely nudge him past 270 shows how difficult a case this is to make.