Slightly revised on 11-28-08 to reflect a 7 point Obama win, as he is now up 6.7 points in votes counted as of now, so 7 is closer than 6 (which is what I used when I first posted this entry)
In that endless post, I predicted results for a ton of states. Needless to say, that post was made well before the selections of Biden and Palin, the worst of the financial crisis and before the campaign had begun in earnest.
That post assumed my prediction of an 8-point Obama win. As of this writing he is up by 6.7 points (votes are still being counted even though the winner obviously is not in doubt), so my state-by-state predictions are valid-- had Obama won by only 1, my state-by-state predictions of Obama's final margin would have been worthless.
Once again, here is my terminology. In 2004, Bush won by 3. A state that went for Bush by 8 points is red + 5: 5 points more GOP than the nation. A state that went for Kerry by 10 is blue + 13.
Obama is (as of this writing) up by about 6.7 points, 52.8 to 46.1. I am rounding up to 7 points. So a state that went for Obama by 3 is red +4.
I note one big embarrassment before I begin. I did not even consider the possibility of Indiana being a swing state. It has historically been very red, being red + 18 in 2004. I just didn't think Illinois bordering Indiana was worth anything remotely like that. I was, needless to say, badly wrong, as Indiana ended up going for Obama by less than one point, thus moving Indiana to red + 6.
1)Florida: 27 electoral votes: 2004: Red + 2.5.
My June prediction: Obama by 3 (thus red +5)
Actual results: Obama by 2 (thus red + 5)That went well! I realized that Florida would be harder than other swing states due to a relative lack of Obama's best demographic groups (upscale professional whites and black voters). Frankly, everyone else that looked at this election did as well. I may have hit Florida almost perfectly, but its difficult to take too too much credit for seeing what everyone else saw.
Florida went exactly as expected. Both campaigns fought hard, McCain took a small lead when he was up a touch nationally, Obama regained the lead as he pushed ahead in the national polls, and maintained a small but very consistent lead in polling from Mid-October on. Florida went EXACTLY as the polling expected, in roaring high turnout.
2) Ohio: 20 Electoral Votes: 2004 Red + 2.5
My June prediction: Obama by 3 (thus red +5)
Actual results: Obama by 4 (thus red + 3)
Oops. Ohio went better for Obama than I expected. I was frankly mildly concerned about race here. Like Florida, Ohio was expected to be mildly difficult for Obama, was, and went almost exactly as the polls said it would. Obama was consistently up low to mid single digits in Ohio. I suppose McCain outperformed his polls by a tiny bit, but not by enough to begin to matter. My Ohio prediction was conventional wisdom 101, and we were all a tad off.
3) Pennsylvania 21 Electoral Votes: 2004: Blue + 5.
My June prediction: Obama by 14 (thus Blue + 6)
Actual results: Obama by 11 (thus blue + 4)
I predicted in June that McCain would abandon Pennsylvania fairly early on, because its blue, and expensive. I was hugely wrong here. McCain abandoned Michigan early, but fought like crazy in Pa, sending Palin in every 20 minutes and very often appearing himself. The McCain campaign spent huge amounts of money here. And what did they get for their troubles? Nothing. It was blue + 5 in 2004, and after all the campaigning and phone calls and events and tv ads, it was blue + 4 in 2008! I think its fair to say I was very right in June and Team McCain was wrong. Pa was just a bridge too far. I note that I predicted that, "the polling will favor Obama solidly here." On that prediction I was very very right. Pennsylvania polls consistently put Obama up by low to mid double digits until the final week of the campaign, at which time the McCain efforts showed up in the polls, where Obama led in the mid to high single digits. But the final result, in low double digits, was entirely consistent with my June prediction, and with the polls. The Pennsylvania gambit was a good idea by the McCain campaign as of mid-October, as they just had no way to win without it. Not that they really could have won with it, but when you have lemons, you make lemonade.
4) Michigan: 17 Electoral Votes: 2004: Blue + 5.
My June prediction: Obama by 11 (thus blue + 3)
Actual results: Obama by 16 (thus blue + 9)
Oops. I blew this prediction, from start to finish. So much for a race-related problem. I predicted McCain would make a big push in Michigan, and would more than likely abandon it early in the Fall. Well, I was right about that, yet missed the victory margin by 5 points. Or, put another way, I missed the skew towards the democrats by 6 points. That's a ton in a seemingly predictable swing state! In the end, I thought Obama would win Michigan by more than he won the nation by, and he did.
My reasoning was that the Michigan economy had been in the crapper for ages, and maybe there McCain could say hey, try something new here, your democratic governor has messed things up. And he had polled well in Michigan. But literally days after McCain pulled out he was down mid double digits in the polls, instead of high single digits. And that's where the vote ended up. I still think Michigan was a better choice than Pennsylvania. I may have missed my prediction, but the McCain campaign missed the choice.
5) Virginia: 13 Electoral Votes: 2004: Red + 5.
My June prediction: Obama by 7 (thus red +1)
Actual results: Obama by 6 (thus red + 1)
Ah, Virginia. The Old Dominion. There's very little I like better in life than being hugely right. And, well, when it comes to Virginia I was right start to finish. Spectacularly, presciently right.
Not only was I very right, but Virginia mattered hugely in the campaign because of the electoral math. As I have said in other posts (and the Obama campaign said early on), if Obama carried every Kerry state (which he was favored to and did) (252), and Iowa (which he was favored to and did) (7), flipping Virginia alone (13) would win him the election, even if he lost in Ohio, Florida, Colorado, and everywhere else. This nasty fact caused 2-months of abject panic by the McCain campaign, once they realized how serious Obama was about Virginia, and, in my opinion, was by far the largest reason for the McCain campaign's seemingly strange insistence on campaigning hard in Pennsylvania and somewhat in Iowa. Virginia was the 13 electoral votes that terrified Team McCain.
Unlike Florida, where I was just as right, but really so were a lot of other people, I broke against the great weight of conventional wisdom when I wrote my June post. At that time, most pundits agreed that Colorado was a legitimate swing state, but that Virginia was a bridge too far. The conventional wisdom was that Obama would play there, then back off, and McCain would ultimately win it, even if he lost the election. Not! In the end, it was demographics. As I noted, "Virginia has been moving towards the democrats almost by the hour for the last 12 years." Specifically, as I noted, "Virginia was Red + 11 in 1996, and Red + 8 in 2000, so it has trended blue for a while now. It moved 3 points towards blue from 1996-2000, 3 more points from 2000 to 2004, and could easily have moved that much or more since." And that's what happened. In conjunction with a monster Obama campaign effort and monster black turnout, Virginia shifted 4 more points towards blue. Why?
To put it simply, a huge chunk of voters who vote basically like Maryland (which went for Obama by 25 points) has moved in, in slow motion, since about 1995. All I predicted, in the end, was that this demographic trend would continue, that Obama would play hard for Virginia, and that any race effect in Appalachia would be cancelled out by his huge popularity in northern Virginia, where the people that vote like Maryland moved. And, as everyone else predicted, black voters would turn out in breathtaking numbers.
I also predicted that Mark Warner would win his Senate seat huge, possibly by as many as 20. That counts as a miss; he won by 31. One former governor beat another former governor by 31 points. I wonder if that's ever happened before in a Senate race.
In conclusion, Virginia went precisely as I thought it would for precisely the reasons I thought it would. When I wrote the June post, the polling average nationwide had Obama up by 3.9 points, and in Virginia, McCain was up by 1.3 (or red + 5.2). I looked past the polls, saw the demographics, predicted the all-out effort by Obama, and got the final result almost exactly right.
6) Missouri: 11 Electoral Votes: 2004: Red + 4.
My June prediction: Obama by 2 (thus red +6)
Actual results: McCain by less than 1 (thus red + 7)
I predicted Missouri would be tough for Obama, but he'd win it if he won the popular vote comfortably. So did everyone else. Everyone else predicted that Missouri has moved somewhat to the red side and is no longer a true bellwether. I even got the partisan hue, red + 7, almost exactly right. We were all correct, except that he lost it by a hair instead of winning it by a hair. Moving on.
7) Minnesota: 10 Electoral Votes: 2004: Blue + 6.
My June prediction: Obama by 14 (thus blue + 6)
Actual results: Obama by 10 (thus blue + 3)
A 3-point miss. No big. I predicted that McCain would abandon it. He largely did. The GOP held their convention there, but they still realized the inevitable. Interesting that Minnesota continues to move red. I have no idea why, I need to learn that.
8) Wisconsin: 10 Electoral Votes: 2004: blue + 3.
My June prediction: Obama by 13 (thus blue + 5)
Actual results: Obama by 13 (thus blue + 6)
Not bad I'd say! All I predicted here (as in many other states) was that the partisan hue of the state from 2004 would continue. Obama always polled well in Wisconsin. He whipped Hillary in the primary there. Anyway.
9) Colorado: 9 Electoral votes: 2004: Red + 2.
My June prediction: Obama by 10 (thus blue + 2)
Actual results: Obama by 9 (thus blue + 2)!!!!!!!
Bingo. WOO HOO. As with Virginia, I was hugely right here. Now unlike Virginia, everyone and their grandmother predicted Colorado was a swing state in 2008. But most people thought it would be much closer. Here, as in Virginia, I correctly predicted that demographic trends moving the state towards the blue side of the ledger would continue, and I was very right. Colorado was blue +2. Colorado was blue +2. Colorado's been red for a long long time, for it to be blue + 2 is astounding. According to the exit polls, Hispanics in Colorado went huge for Obama, 61-38, but that's actually a tad lower than the national results. Black voters made up only 4% of the voters. What happened in Colorado was, incredibly, Obama won the white vote 50-48. That is absolutely incredible. By way of comparison, the white vote in deep deep blue New York was only 52-46 Obama. Whites in Colorado, a traditionally red state, voted at nearly as high a rate for Obama as in blue blue blue New York. WOW.
Colorado was a part of a larger story, where the west moved hugely blue in 2008. Even the 3 west coast states, all Kerry states, moved blue. And Colorado, Nevada and New Mexico moved sharply blue.
As I wrote back in June, in 1996, Colorado was Red + 10, in 2000 it was red + 9.5, but in 2004 it shot up to red + 2, a 7 point move towards blue (and with the same GOP candidate, Bush). I figured it would go 4 more points towards blue, and it did, exactly 4 more points.
10) Iowa: 7 Electoral votes: 2004: Blue + 2.
My June prediction: Obama by 13 (thus blue + 5)
Actual results: Obama by 9 (thus blue + 2)
Ah well, I suppose being three points off on the partisan hue isn't that bad, but I feel like I was wrong here. I predicted that McCain would abandon Iowa because of his brave opposition to ethanol subsidies. I was badly wrong about that-- he even visited in the weekend before the election. I wonder if the somewhat vigorous McCain campaign in Iowa accounted for the partisan hue remaining at blue + 2 rather than than my predicted move to blue + 5. Or I could have just been plain wrong. No way to tell. But I'm taking Iowa on the chin because I was so confident that ethanol would play a gigantic role here, and it looks as if it really didn't.
11) Arkansas: 6 Electoral Votes: 2004: Red + 6
My June prediction: McCain by 5 (thus red + 13)
Actual results: McCain by 20 (thus red + 27)
Well, if you're going to be wrong, you may as well totally blow it! That's really embarrassing. At the time I wrote the June post, McCain was up 15-20 points on Obama in the polls. I thought that would drastically close. Obviously not. Obama stank up much of the south, losing former swing southern states by gigantic margins. Obviously race played a real factor, as Kerry certainly wasn't popular in Arkansas, and he hugely outperformed Obama! Instead of holding near red + 6, or maybe ticking up to red + 9, Arkansas shot up to red + an astounding 27. I knew Arkansas wouldn't really be a swing state, but being that wrong on the margin of victory is, well, not ideal.
12) New Mexico: 5 Electoral votes: 2004: blue + 2.
My June prediction: Obama by 4 (thus red + 4)
Actual results: Obama by 15 (thus blue + 8)!!
Western Oops, part I. I missed the partisan color of New Mexico, a swing state, by a mere 14 points!!! Oops!! You'll forgive me if I hide under the covers! At least I knew I had no real idea about how New Mexico would end up.
No one else as of June predicted anything like a 15 point Obama landslide in New Mexico. I badly underestimated how well Hispanics would rally to Obama. According to the exit polls, the Hispanic vote was 41 % of the vote in New Mexico! Obama won it 69-30!!! The immigration debate of 2007 did huge harm to the GOP brand among Hispanics, particularly Hispanics living near the Mexican border. This harm has the potential to be generational, like the harm done to the GOP brand by its opposition to civil rights in recent decades.
I said in June I had no confidence in my New Mexico predictions. Needless to say, to paraphrase Churchill, my utter lack of confidence was well founded. For what its worth, the conventional wisdom predicted a fairly close race in New Mexico until about the last 2 weeks, and even then almost no one was predicting the 15-point massacre handed out to McCain and Palin.
13) Nevada: 5 Electoral Votes: 2004: Blue + 0!
My June prediction: Obama by 4 (thus red + 4)
Actual results: Obama by 12 (thus blue + 5)!
Western Oops, part II. Another Las Vegas sized MISS. See New Mexico, above. The west and the south were the dramatically changed regions in 2008, with much of the south moving hugely towards McCain (more on that in a future post) and some of the west moving hugely to Obama. I thought Nevada would be very tough for Obama, and was, obviously, spectacularly wrong. Again, no one predicted anything like a 12-point win in Nevada, not even in the last few days of the campaign. But just because no one else had a clue doesn't mean I shouldn't have.
One other note on Nevada. Hispanics went for Obama by a staggering 76-22. But they were only 15% of the vote (again all this is according to the exit polls). So these 15% of voters gave Obama about a 5-point statewide lead. Black voters made up 10% of the Nevada electorate (I'm surprised it was that high!) and went 94-5 for Obama, thus giving him 4.5 additional statewide points. White voters only went for McCain 53-45, which was a touch closer than in the rest of the country. In the US as a whole, white voters went for McCain by 55-43. That Nevada whites would be bluer than the rest of America tells us something about the GOP brand out west.
14) New Hampshire: 4 Electoral Votes, 2004: Blue + 4
My June prediction: Obama by 7 (thus red + 1)
Actual results: Obama by 9 (thus blue + 2)
I was correct that McCain competed there till the bitter end. I slightly underestimated just how badly tarnished the GOP brand was in New Hampshire, a state which got to know McCain better than any other state except Arizona.
Here are the non-swing state predictions I made.
A) California: 55 Electoral votes: 2004: Blue + approximately 14.
My June prediction: Obama by 15 (thus blue + 7)
Actual results: Obama by 24! (thus blue + 17!)
To miss by that much in a state so huge is embarrassing. Its a stunning monster landslide! Why was I so badly wrong?
Apprently a fair bit of the GOP base stayed home. Even that doesn't account for the huge margin by which I was wrong. I was predicting a 7-point change in the partisan hue, which was daring. Still....
B) Texas: 34 Electoral votes: 2004: red+ 20
My June prediction: McCain by 8 (red + 16)
Actual results: McCain by 11 (red + 18)
Plenty close enough. I'm declaring victory and moving on.
C) New York: 31 Electoral votes: 2004: Blue + 22.
My June prediction: Obama by 26 (thus blue + 18)
Actual results: Obama by 25! (thus blue + 18!)
Now THAT'S what I call knowing your home state! Predicting a big landslide in New York for Obama was child's play. That I hit the partisan hue exactly is satisfying.
D) Illinois: 21 Electoral votes: 2004 Blue + 14.
My June prediction: Obama by 30 (thus blue + 22)
Actual results: Obama by 25 (thus blue + 18)
A slight miss. I guess I overestimated the home-state effect. Put another way, there are always a certain percentage that just are NOT going to vote for you.
E) North Carolina: 15 electoral votes: 2004 red +9
My June prediction: McCain by 3 (thus red + 11)
Actual results: Obama by less than 1 (red +7)
Wow did I badly underestimate how much North Carolina had changed! In June, I advocated Obama going after North Carolina, but didn't think he'd get there. I guess I've been away from North Carolina since 1996, and don't really know it anymore. Still, do note that it was red + 7. Still a long way from being a national bellweather, unlike Virginia, which is at this instant in time the closest thing we have to a national bellweather. Sorry Missouri. No love for you, Ohio.
F) Georgia: 15 electoral votes: 2004 red + 14
My June Prediction: McCain by 6 (thus red + 14)Actual results: McCain by 5 (thus red + 12)
Not too shabby. Still, for Georgia to be red + that much means it is unlikely to go blue anytime soon, absent a big national landslide.
G) New Jersey: 15 electoral votes: 2004 blue + 9
My June prediction: Obama by 15 (thus blue +7)
Actual results: Obama by 15 (thus blue + 8)
New Jersey is right next door to my New York, and simply hasn't changed much in recent years. It wasn't too tough to predict, though of course getting the partisan hue within one point and change is satisfying. Not difficult, mind you, but still satisfying. I knew McCain wouldn't play for New Jersey, despite saying he would on more than one occasion. The McCain campaign may at times have been stupid; competing for New Jersey, requiring the expensive New York tv market, would have been suicidal. They weren't suicidal.
No comments:
Post a Comment