Gosh, I love this stuff. Sorta addicted.
Some people -- like me -- just love sorting through political data. I get a thrill looking for trends, trying to smooth out outliers, and in the best of all possible worlds, accurately predicting future races. Every morning, before I get out of bed, I reach over to my laptop and pull up Rasmussen's daily, Gallup's daily, Hedgehog, Right Pundits, RCP averages, and 538. My daily fix.
I've met quite a few political "junquies" who spend considerable amounts of their waking hours devising formulae for predicting upcoming elections. They're sick puppies -- really! -- who can't quit thinking statistics. They're the ones who use words like autocorrelation and heteroskedasticity. Huh? Predictably, their statistical myopia obscures emotions and passions within the populace. They don't see it and thus, don't understand it. Blinded by the world of numbers, they're useless ... almost.
Any formula must start with the voter's inertia, or the unwillingness, in the short term, to admit they've been duped or made a mistake, and in the long term, to change party affiliation or sympathy. People don't like to change, in spite of Obama's rhetoric to the contrary.
Also, the formula must somehow quantify how different events or happenings effect the natural inertia of voters. This isn't easy. Politicians often work under the radar of ordinary voters. In normal years, voters aren't very motivated to gain information. So ignorance, in itself, has to be a factor contributing to inertia.
So far, inertia + ignorance ...
When voters are well informed by the MSM as well as alternative media sources such as Fox news and talk radio, their skepticism rises quickly. Currently, the health care debate is an example of voter ignorance changing to knowledge. Ordinary people are actually trying to read the opaque bill. They're watching Glenn Beck. Listening to Rush. And the upshot of this voter "education" is passion.
So, so far, inertia + (ignorance - passion) ...
Most politicos love to add economic data to the formula. Its clean. Easy to quantify. Accessible. And it makes the formula look so academic! As the above referenced article puts it, "The motivations for including growth in per capita GDP, real taxes, and real transfer payments in our regressions of course is to measure the impact of economic conditions on election outcomes ..." Did you see that "of course?" What a snit.
Here's the interesting stuff.
Every ONE percentage point increase in per capita GDP increases the presidential party's share of by 1.47 percentage points.
Every ONE percentage point increase in real transfer payments adds 0.12% to Obama's increase.
Every ONE percentage point increase in per capita federal taxes subtracts 0.36% from Obama's midterm election results.
And then he fudges. He says, in reality, that the impact of growth in output and taxes to be larger than the impact o government transfers. Another interesting tidbit is that taxes and benefits at a local level do not have any impact on how voter's react in federal midterm elections. Really?
So here's the formula, with X as the leading coefficient or variable:
Inertia + X(information efficacy)(1.47GNPdelta)(0.12transferpaymentsdelta)(-0.36federaltaxdelta)
Pretty cool, huh?

