Why Romney will win the presidency tomorrow

This post is the final article in a three-part series analyzing likely outcomes of Tuesday's Presidential Election. It is not an endorsement of any candidate or party.

On Friday, we looked at reasons why President Obama might win re-election in tomorrow's presidential election. In particular, we examined the fact that Romney needs to win a high number of states that Obama carried in 2008, Romney's disadvantage in state polls, academic models, like FiveThirtyEight.com and prediction markets like InTrade.

These four factors might seem insurmountable, but in reality, there is a fair amount of overlap. Models like FiveThirtyEight are constructed entirely on the basis of state polls, and Prediction Markets are largely predicated on these polls as well. That means if the state polls are wrong, almost all of the president's advantage melts away.

The bad news for Romney supporters is that polls can typically predict the outcome of an election within a couple percentage points. For example, polling averages nailed the final results of Scott Walker's victory in June and also correctly predicted Obama's popular vote victory in 2008. That polls are frequently accurate within two percentage points might be a note of consolation to Romney fans. After all, in a race as close as this, if he overperforms the polls by two percent, it might very well be enough to carry him to victory in critical battleground states. Still, this is of limited value as the polls could just as Romney could just as easily underperform the polls. As we learned from Disney's The Mighty Ducks:

Coach Gordon Bombay: A quarter of an inch this way and it would have gone in. A quarter of an inch, Charlie.

Charlie: Yeah, but a quarter inch the other way, and you'd have missed completely.
(I suspect the Badger Catholic is the ONLY news source that features Mighty Ducks quotes into its election analysis.)

So if you want to feel confident about Romney's chances, we're going to want to find some way to prove that the polls are somehow systematically flawed. Is that possible? It just might be.

Polls Biased?
Many of the polls showing Obama in the lead--particularly at the state level--are predicting that Democrats vote in greater numbers than Republicans by somewhere between five and ten percent. That would mean that Democrat turnout matches or exceeds Democrat turnout in 2008. That seems unlikely, given that 2008 was a wave election for the Democrats, and conservative enthusiasm for ending the Obama presidency appears high.

One theory suggests that perhaps polls, particularly in battleground states, are marred by a systematic (as opposed to an ideological) bias. Under this theory, many polling firms--particularly those run by mainstream media outlets--are facing budget constraints that require them to cut corners on their polls. As a result, they are failing to adequately distinguish between registered voters and likely voters. But among registered voters, Republicans are much more likely to vote. So the polls might be finding a Democratic advantage among registered voters who won't actually cast ballots. If the Democrats' advantage in voter turnout is neutralized to less than five points, Romney polls even or ahead in most polls. This neutralization is precisely what a number of pollsters--including long established and widely established Rasmussen, Gallup and Pew--expect.

"This theory is very nice, Steve," you might be saying. "But is there any empirical evidence supporting it?"

Excellent question, and I'm glad you asked. As a matter of fact there is. For starters, pollsters are reporting that in a number of battleground states, approximately 40 percent of voters have cast their ballots early. But these numbers can be verified, and in reality, the early voting rate has been significantly lower. As Democrats typically enjoy a strong advantage in early voting, this could negate Obama's predicted turnout advantage to fewer than five percentage points.

More on Turnout
But that's not the only reason Romney might win the turnout battle. As we just mentioned, Democrats usually win handily among early voters, leaving Republicans to play catchup on Election Day. But early data suggests that Republicans might have closed that gap. In states that register voters by political party, the number of registered Democrats who have voted early has either declined or increased much more slowly than the number of registered Republicans who have voted early. Similarly, early voter turnout in Democrat strongholds (like Milwaukee) has increased at a much slower rate than Republican strongholds (like the Milwaukee suburbs).

This advantage could be fleeting as it's unclear whether early voters of either party would have voted on Election Day. But for the Romney campaign, early results are certainly preferable to beginning with a heavy pre-election deficit to make up. According to GOP insiders, the Romney campaign is running an incredible get-out-the-vote (GOTV) operation--one exceeding even the vaunted Bush ground game of 2004. If Romney wins in states like Wisconsin, Iowa and Ohio, it will almost certainly be driven by strong GOTV.

The top line "horse race" polling results aren't the only thing to look at. We should also break down the results. Most polls--even those that show Obama with a meaningful lead--show Romney winning among independents by a healthy margin of between 5 and 15 percent. If this lead holds, it will be very difficult for Obama to win.

Other key indicators that favor Romney include crossover support (voters supporting the candidate of the opposite party), voter enthusiasm, personal approval, and preference for the candidates' economic policies.

Late-Breaking Developments
While polls can provide an excellent snapshot of the electorate in any given moment, they do tend to lag a bit. Obama's recent surge in the polls is most likely reflecting voters' perceived favorable response to Hurricane Sandy. That bounce is likely to fade before Tuesday--even if the polls don't reflect it. As New Yorkers continue to suffer, the initial favorable reaction to Obama's response might reverse itself and lead to an anti-Obama surge. Other recent developments, including a lackluster jobs report from Friday and new revelations about the president's handling of the Libya crisis could also lead to a late break for Romney.

The Incumbent Rule
We might also examine that challengers tend to overperform their polling numbers. Usually, it's not enough to matter, but--again--in an election this close it might make all the difference. There's no guarantee that late-deciders will break decisively for Romney. But it's almost certain that they won't break for Obama. Romney's worst-case scenario is a 50-50 split.

Will these reasons be enough to help Romney complete the upset? It's anybody's guess. I'll chime in with my prediction tomorrow morning, and we should know for sure tomorrow night.


Greg (DI and DII fan...DII was mediocre) said...

quack.....quack...quack..quack. QUACK QUACK QUACK! Goooooooo DUCKS!

Angie said...

Really enjoyed all three of these posts! Thank you.