As someone who follows voters a lot less carefully than John does, I'm still fascinated by this, and left with a whole bunch of questions. Note: it's possible that some of these have been studied and answered. But I'll lay them out there, and see if someone knows more than I do -- or if not perhaps someone might think about finding out the answers.
1. One interesting thing is that unlike, say, whether scary weapons were found in Iraq, perceptions of the economy could be based on personal experience as much as they are based on national information. Do we know whether there's any reason to believe that partisans actually experience a different economy? After all, there are both geographic and other demographic differences between Democrats and Republicans (I'm pretty sure the answer is no, that this is entirely driven by partisanship, and that it flipped some time in early 2009. But is there anything to the other possibility at all?).
2. Do we know to what extent there perceptions are driven by pure partisanship vs. partisan sources of information? That is, if Fox constantly says that the economy stinks when Obama is president, but not when Bush is president, then anyone who watches will believe that...and since more Republicans than Democrats watch Fox, it will produce a national split. On the other hand, it might be be that Republicans automatically switch to believing the economy has tanked as soon as a Democrat becomes president, and vice versa.
3. In the fourteen months or so covered by the chart John reprints, Democrats seem to be much open to changing their opinions than Republicans; the total range for the Democratic line is 54 points, while the Republican line has a range of 31 points. The implication is that Democrats are responding to news events, while Republicans are just certain that the economy is lousy no matter what happens (or, perhaps, that they're hearing bad new no matter what happens). Does that flip when a Republican is in the White House? That is, is it caused by partisan differences, or electoral context?
4. Does partisan polarization about the state of the economy also affect perceptions of personal economic circumstances?
5. And finally: are these just survey answers, or do they reflect some underlying beliefs that might actually translate into behavior? That is, do the perceptions of many Republicans that the economy is horrible actually lead them to behave as if the economy is horrible?