Since this thread desperately needs rescuing, I'm going to talk a bit more about how stupid voters are.
We (geniuses) have known for a long time that one of the primary determinants of a president's chances of getting re-elected (or plain elected, if he's the challenger) is the real disposable income growth that occurs in the country during his four years in office. This is a measure of the amount of income that becomes available to workers after inflation.
Now, voters are stupid about this on several levels. First off, they believe the president has control over the economy - this is a dubious prospect at best. Second, however, a rational voter would (probably) look at RDI growth over all four years of the president's term. It's possible that they would attribute the early years to the previous president's policy, but actual questions don't seem to suggest this is true - they don't even think about that.
So what do they do instead? Well, really they only seem to care about RDI growth in the election year. If the election year went well, they vote to re-elect. If the election year went badly, they vote to kick the bum out. They don't even think about the past three years.
And do presidents know this? Well, we're not sure. It's looking like they might; average historical RDI growth in election years is nearly double that in non-election years.
Voters might be cleverer than we suspect, too. We're going through cached polls as far back as the 1930s to try to spot questions asking about 'election year gimmicks' and related behaviors.
Interesting stuff. The general term for it is 'voter myopia'. We've even managed to replicate it in an experimental paradigm, using small amounts of money and an automated allocator the player can choose to keep or reject.