Top Republicans try to distract attension from their lackluster campaign to blame their electoral woes on the economy, claiming a structural relationship between economic anxiety and polling.
Below the fold, I explore whether or not this relationship actually exists.
*****Cross-Posted at Stochastic Democracy*****
Talking about separating the economy from the "Palin Effect", Jay Cost writes
I'm not saying this is what's really going on. I think this is a reasonable alternative hypothesis to the reader's "Palin Effect" theory. Complicating matters even more is the fact that the two ideas are not mutually exclusive. Some voters might be moving from McCain because of Palin. Other voters might be moving away from McCain because of the economy[emphasis mine], and that's eroding Palin's numbers. There might be other explanations, too - so overall we have a tricky situation on our hands.
Cost implicitly assumes that Economic Anxiety helps the Democrats. I've heard this frequently from the bloggers, and it was even suggested by McCain's campaign manager.
But is it true? To make a long story short, most likely not.
The statement "Economic Anxiety pushes voters toward Obama in the Polls" can be mathematically formalized and tested in many theoretically plausible ways*. And unfortunately, none of them that I could think of are true. (Of course, I'm not infallible, feel free to leave ideas in the comments)
The Economy has gotten worse since August, and Obama has improved his position since August. So I understand why a lot of people have assumed causation between the two, but by that logic, Pirates Cause Global Waming. If someone goes ahead and to tries to fit an ARIMA model, they'll find that nearly every economic variable imaginable has an insignificant effect on day to day polling changes.
Smoothed Graph of Obama's Vote Share
Graph of Daily Closing Prices of the Dow Jones Industrial Average(INDU)
Daily Close of the VIX volatility index
I don't want to get too far outside my comfort zone, but to me, the graph's long and near-linear increase is more consistent with McCain running a really bad campaign more than anything else.
*For example, one of the more elegant models I tested was that polls follow a random walk biased by economic anxiety
Where e_t is a random white noise process. This model possesses many awesome properties(It's a martingale by the Efficient Market Hypothesis for most Economic Indicators). But Alas, it was not significant no matter what variable I used to track Economic Anxiety.
*****Cross-Posted at Stochastic Democracy*****