Sunday, August 26, 2012

Arizona Is (Probably) Not a Swing State

Two polls released this week show the presidential race essentially tied in Arizona, with one giving Barack Obama a lead of two percentage points and the other giving Mitt Romney a lead of two percentage points. The Obama campaign has suggested it will compete in Arizona, a state it largely ignored in 2008 because it was John McCain’s home.
But is Arizona really a swing state?
Let me remind you about how I use the term “swing state” here at FiveThirtyEight. When I employ the term, I mean a state that could swing the outcome of the election. That is, if the state changed hands, the victor in the Electoral College would change as well.
The most rigorous way to define this is to sort the states in order of the most Democratic to the least Democratic, or most Republican to least Republican. Then count up the number of votes the candidate accumulates as he wins successively more difficult states. The state that provides him with the 270th electoral vote, clinching an Electoral College majority, is the swingiest state — the specific term I use for it is the “tipping point state.”
From Barack Obama’s perspective in 2008, for instance, his easiest three electoral votes were in the District of Columbia. The next-easiest were the four electoral votes in Hawaii, giving him seven total. Repeat this process and you find that Colorado was the tipping point state in 2008, putting him over the top with 278 electoral votes. (Although, winning Iowa but not Colorado would have sufficed to give Mr. Obama 269 electoral votes, an exact tie in the Electoral College.)
Note that the swing states are not necessarily the closest states. In fact, they won’t be, unless the election itself is very close. In 2008, the closest states were Indiana, North Carolina and Missouri. However, those states were superfluous for Mr. Obama — North Carolina, for instance, brought him to 365 electoral votes, nearly 100 more than he needed.
Likewise in 1984, when Walter Mondale was clobbered by Ronald Reagan, the closest states were Minnesota and Massachusetts. But Michigan was the tipping point state; its 20 electoral votes were the ones that put Mr. Reagan over the top.
If you’re not going through all this work to sort the states in order — and it does require some work, since you need to project the outcome in every state as well as account carefully for the uncertainty in the forecast — a reasonably good alternative is just to compare the average of state polls to the average of national polls. The closer the state is to the national average, the more likely it is to play a decisive role in the election.
Right now, for instance, Mr. Obama leads Mr. Romney by about four percentage points nationally, so the swingiest states are those where Mr. Obama also holds about a four point lead. In Ohio, for instance, Mr. Obama leads Mr. Romney by about five percentage points in recent surveys, so it is (unsurprisingly) a swing state again this year.
How about Arizona? The two most recent polls should show a rough tie there — but since Mr. Obama leads in national polls, it’s still about four points away from the tipping point.
And those polls may have been modest flukes. A longer-term average has Mr. Romney three percentage points ahead there, meaning that Arizona leans a net of seven points more Republican than the average of national surveys.
Arizona showing a seven-point Republican lean relative to national numbers would be fairly normal. That’s about where it was from 1992 through 2004, for instance. The gap was wider in 2008, but presumably because its native son, Mr. McCain, was on the Republican ticket.
None of this is to say that Mr. Obama couldn’t win Arizona — he certainly could. Bill Clinton won Arizona in 1996 when he won the election by about eight percentage points nationally. If Mr. Obama won by a similar margin, he’d be at least even-money to pick up the state as well.
But if he does win Arizona it will probably be superfluous, since in all likelihood he’ll already have won states like Ohio, Colorado and Virginia that are closer to the tipping point.
The situation is analogous to Mitt Romney winning a state like New Jersey, which is also plausible if Mr. Romney wins the election by several points nationally. But that doesn’t make New Jersey a swing state — it’s just extra spoils for the winner.
With all that said, there is a lot of uncertainty this far out from the election. The ordering of the states is usually fairly consistent from year to year, especially when there is an incumbent running for re-election, but they can be scrambled to some extent by local issues, long-term demographic trends or the amount of attention that the campaigns focus on them. It’s not impossible that Mr. Obama could pursue a “Western strategy,” adding Arizona while holding Colorado, Nevada, New Mexico and the Pacific Coast states. That would give him a lot of slack elsewhere.
But the benchmark for whether these possibilities are realistic and relevant to the Electoral College equation — and whether Arizona should really be thought of as a swing state — will be if Mr. Obama starts to poll consistently as well in Arizona as he does in states like Ohio. The state is worth keeping an eye on, but we haven’t seen that yet.


Share your thoughts.
    • Mark Johnson
    • Virginia
    I respect Nate's work in general but this line of reasoning is a real nothing burger. It has no usable point and the redefinition of "swing" is a none starter. (And Bob H. no, "purple" is not the same as swing and even if it was "swing" was used long before "purple" was.) The only scenario where this model would be of any benefit would be if one knew in advance what order the states would be in. State voting is contingent on many different factors and the order of the states on the Nate's Swing Graph will shift accordingly.
      • Bob H.
      • Ohio
      Silvers's analysis is good and his definition of swing state is correct. We already have a term for states which are simply close: "purple" states. Nobody would buy that a purple state with a small number of electoral votes is likely to be relevant. Similarly, only those states near the 269 mark are likely to decide the election.

      His example of North Carolina illustrates this well; nobody really cared nationally that Obama won NC since it was overkill.

      If we believe the premise that the continuum of the states is relatively stable (with exceptions for candidates' home states), then the conclusion follows. Maybe we could see some more evidence that the ordering of the states is stable.
        • Watermark
        • MS
        "Let me remind you about how I use the term “swing state” here at FiveThirtyEight. When I employ the term, I mean a state that could swing the outcome of the election. That is, if the state changed hands, the victor in the Electoral College would change as well."

        That is not the way most people use that word, though. You can't just redefine a word away from common usage and chastise others for not conforming to your redefinition.
          • Myron
          • Fayetteville, NC
          I guess it's OK that Nate comes up with his own definition of a swing state, but I'm curious as to what's the point.

          From a political perspective, what matters is that a state is important enough electorally and unclear enough politically to where both candidates have to spend their two most valuable resources there -- time and money. How swing states as a whole are managed is what affects the outcome -- not a single-minded focus on what state is deemed to be THEE swing state.

          States like NC, Arizona, Indiana, and Virginia may not be thee swing state, but they could be collectively decisive for either candidate -- in particular the GOP candidate whose strategy relies on a solid south.

          Obama's "spread the map" strategy is precisely how he defeated McCain. If he had narrowed his focus to the traditional Democratic path to victory, yes he might have still won (since he did win Ohio, Florida, etc.) but with less of a cushion.

          This year, we see he needs that cushion.
            • JDK
            • MD
            Problem 1: making a new definition of "swing state"
            Problem 2: using pretty pseudo-scientific ("scientistic") method. By putting states in rank order you are in essence asserting a claim that +9.5 is meaningfully different than +9.6. etc. why? Also things are different in 2012, aren't they. President probably has better chance of winning IN and NE2 than winning NC and VA, and will have to compete much harder in NH and NV, just for starters. The only way that this approach makes any sense is to start with releasing the factors in your final model. And then updating based on new demographic data (you used 2000 census instead of ACS and now there is 2010 census) and the 2010 elections, plus refined nearest neighbor. Otherwise your just being a second rate political commentator with a bias AND a data charlatan or worse a guy in Vegas who thinks he has a "system".

            Step it up or step out.
              • zoeae
              • Montclair NJ
              I think it is more useful to think in terms of a group of "swing" states, using the Silver definition. Before seeing the electoral map above, I was working with 3 states - OH, PA and FL -- and assuming that whoever wins will have to take 2 of these 3. The Western strategy reflected above shows how Obama might win even while losing OH and FL. But even here, I see a group of swing states -- Obama would have to take AZ, IA and PA in order to lose both OH and FL and still win. Of course, I don't see Obama winning AZ, so I still think the campaign should pour enormous resources into OH, PA and FL.
                • Glenn Doty
                • Columbia, SC

                I think there's less than a 3% chance that the Republicans will win PA. It's simply not a swing state, and it's becoming less "swingy" every year. While it's possible that a republican could win it if they win in a landslide, it's not likely that they'll be able to take PA unless they already win all of the swing states.

                The reds have wasted money campaigning in PA every one of the last 5 election cycles... but they aren't going to win it. They just won't.

                As far as actual battleground states, my map has the following (from west to east): NV, AZ, IA, MO, IN (similar to PA - not a real swing state - but the Obama campaign will contest it if Lugar loses the primary), OH, FL, VA, NC (also similar to PA), and NH.

                Romney is starting far behind, so he's likely to play defense and keep his money in the states he feels he has a chance... I doubt he'll contest any state other than the above listed. He'll have to win NC, AZ, and IN (easy), FL (unlikely), and OH and VA (probably impossible), as well as either winning MO and IA or NV or winning IA, NV, and NH (either option will be very tough).

                That's the battleground as I see it. President Obama will only need 14 EV's from that group - which could be either FL, OH, or NC; or it could be VA, IN, MO, or AZ plus any other state, or it could be NH + NV + IA.

                The last I checked, President Obama was leading in FL, OH, VA, NH, and NV. It's nearly hopeless for Romney, which is great for the U.S.
              • Crissa
              • Santa Cruz, CA
              I would add in, as well, the time that the results were called or were certified; even if Oregon or California were a tipping point state, they would be unlikely to be called at such a point to influence the election, and even though California has a huge media market, I don't think ad-buys there would influence outside the state.
                • Glenn Doty
                • Columbia, SC
                It's important to keep the old definition of "swing state".

                Nate, your definition of "swing state" is not the same as the standard definition. The standard definition is: "a state that may switch from one party to the other in different elections". By that definition, Arizona is a potential swing state this year...

                The reason that is important is that participation in elections is severely depressed in a party-locked state, whether red or blue... and that depressed activity can often lag demographic shifts by several election cycles.

                In 2000, AZ was clearly red, but since that time the adult latino population has increased significantly. Currently ~29.6% of the state is latino, 4.1% of the state is black, and 3.4% of the state is bi-racial. If we assume (correctly) that >90% of the black population and >70% of the latino and bi-racial population vote democratic, then that already equals 26.79% of the vote before the first WASP makes it into the voting both. That means the white people have to break for the republicans by >63%!!!

                Of course, this isn't true by today's voting turnout, because the Latino and black populations, as well as young people and other democratic strongholds, have all understood that AZ was a RED state, and that fact discouraged their participation...

                However, if we correctly call AZ a swing state, then some of these groups that haven't been participating might have hope and start registering, and the republicans lose AZ for a generation.
                  • DirtyLibrul
                  • North Vancouver, BC
                  Couldn't agree more on the definition of "swing state", and how important it is that it remains consistent.

                  Interestingly, for the same reasons you outlined, the title seems to be using the old definition, "Arizona Is (Probably) Not A Swing State", suggests there is more than one swing state. If he uses his new definition it should be "Arizona Is (Probably) Not THE Swing State", since there can be only one (*Highlander reference intentional).

                  I know this is nit-picky and even "gotcha-y", but I find Silver's definition odd and misleading. It's possible he does as well since he switches to "tipping point" halfway through and sticks with it for the duration. That is a much more accurate name.

                  But don't get me wrong, I love me some Nate Silver! Which is why this seems so odd to me.
                • Michael
                • Los Angeles
                It is interesting to read comments by people who question statistics based on their own limited and non-representative samples, e.g., how many bumper stickers they see in their neighborhood. Polls can be wrong, but they are more usually correct. If you are a person who thinks you can beat the odds, I advise you not to visit Las Vegas.
                  • jswiller
                  • highland mills, ny
                  I just read.Dr. Carmona's bio on Wikipedia. The man is unbelievably impressive.
                    • Tom Wolff
                    • East Lansing, MI
                    The list of states, their electoral votes and 2008 margin shows that, for Romney to beat Obama, he will have to win at least one state that Obama won in 2008 by 9% or more. Also, the March 2012 unemployment rates in the following states are all considerably below the national average:

                    Colorado: 7.8%
                    Ohio: 7.5%
                    NH: 5.2%
                    Iowa: 5.2%
                    VA: 5.6%

                    For Romney to win, he'll need to win at least 3 of these states. Seems like a tough task.
                      • Ickey
                      • Evansdale, IA
                      3 states which must necessarily include both OH and VA or else Obama gets to 270.
                      Also, we should point out that Nader isn't running to Obama's left this year, so that far left vote has nowhere else to go but to the President, so those 9+% wins are all actually double-digit wins. So it's an even steeper uphill climb than the raw numbers from '08 would suggest.

                      But always remember that a half-billion in invisible outside unlimited untraceable shadow money can easily buy an election in a lot of states, so nothing's a guarantee.
                    • Sarasota Joe
                    • FL
                    I'm surprised by a few things on the map. You've got IA and ME2 blue but FL, VA, OH red. I think IA and ME2 are redder than any of those, so if those three states all go for Romney, I haver a hard time seeing that IA and ME2 will stay in Obama's column.
                      • bob bob
                      • USA
                      This is your definition of a "swing state" but I think most people just think of it as a state that could go either way (compared to the almost definitely going blue California and the almost definitely going red Utah.)

                      I think "Tipping Point States" (states that could actually make the election go one way or another) is a good term to describe what you're talking about, "Swing State" is a good term to describe a state that appears to be "up in the air" (ie could go one way or another ... assuming it's not a total blowout.)
                        • Jared Hendricks
                        • Bakersfield
                        Arizona will go to Romney and I predict Nevada will too. Why? The LDS population will come out in droves for Romney. It's going to be a close election.
                          • Matt
                          • Minneapolis, MN
                          I think others have covered this already, but doesn't the LDS population already come out in droves for the GOP candidate as it is? Sure, Romney'll probably mobilize a few more than usually vote, but the LDS populations in those states aren't THAT big, and the number of them who wouldn't otherwise vote is fairly small. It's really something that probably made a bigger difference in the primaries that it will in the general.
                        • Lawrence R
                        • Austin, tx
                        I have two questions: Why are we discussing the issue of swing states when the bigger; much bigger story is why the traditional media; CNN, Fox, WSJ and others are not covering the recently announced delegate wins in Iowa, Minnesota, and next Washington state by Ron Paul. Dr. Paul was just in Austin a few days ago and 6000 people showed up to his rally at UT. Why is there no media coverage?
                          • Hayford Peirce
                          • tucson, az.
                          Because he is never, ever going to be even *close* to winning the nomination. He will be totally ignored at the convention. He will never, ever be elected president.

                          Did the "traditional media" pay any attention to Lyndon Laroche?


                          Did the "traditional media" pay any attention to Harold Stassen on the last six of his eight attempts to gain the Republican nomination?


                          That's why the traditional media are today ignoring Ron Paul.
                          • Matt
                          • Minneapolis, MN
                          Wait, wait... You are saying that the "much bigger" story than an actual article about the general election is that Ron Paul has some delegates headed his way? Even though he will not have NEARLY enough delegates in his column to throw the primary into a convention fight (let alone actually take the lead himself)?

                          That's hilarious. Yeah, I'm sure it's the media's fault that people aren't aware that a candidate with no shot still doesn't have a shot, but has a couple of extra delegates in his column.
                          • DirtyLibrul
                          • North Vancouver, BC
                          A) I'm surprised that you, or anyone, still believes Paul has any realistic opportunity, and B) more importantly, why would you expect, or want, today's Republican party to accept Dr Paul? They have really gone off the deep end, and while Dr Paul may have his "interesting" policy stances, the gulf between him and today's GOP is just getting wider.

                          Now if he ran as a Libertarian in 2012 (third party) he would get more than his fair share of attention, probably more than Nader when he ran. So maybe you should make that happen?
                        • Mr Universe
                        • Eugene, Oregon
                        Nate doesn't really give the map at the end of the article enough credit but it's a completely interactive map. You can plug in your own predictions to get to 270 electoral votes. It's a great armchair quarterback type of political bracketing tool. Try it. It's fun.

                          • Terry
                          • AZ
                          I have lived in Scottsdale, AZ since 1990. There is a chance Arizona will go blue this year. Three reasons:
                          1) Mexican-Americans have had it! They are motivated and organizing. A friend's father and uncle (in their 60s) long term permanent residents, recently became U.S. citizens. I said to her, "Now tell them to register to vote." Her response."Oh, they did that on the way home from the ceremony."
                          2) Young women are waking up! One young woman I know voted for Bush twice, blowing off concerns that abortion would be criminalized under republicans, with a, "That can't happen." She laughed when I told her that they (republicans) would go after birth control next. Now, she is appalled that medicare (i.e. the taxpayers) cover Viagra for men (so they can have sex) but Republicans don't want birth control covered for women who's insurance is part of their work compensation, and is often used for reasons other than sex without pro-creation.
                          3) The neo-cons' constant war-mongering. There's a reason young people like Ron Paul, and it's not his views on the gold standard. The last thing they want is a new war in Iran. Don't let the retirement haven designation fool you, Arizona is a young state, a demographic that overwhelmingly favors the Democrats.

                          The democratic base voters are here. The challenge for democrats will be in registering them and getting them to the polls.
                            • rick d785
                            • Oregon
                            Boy do you give me hope. My son in law is Hispanic and a successful executive who travels the world.The one exception he won'take is changing planes in Phoenix when traveling to Latin America.
                            In solidarity I have skipped spring training in Arizona thus the miserable wet season here in Oregon for two years and I miss it.Thank you for your well written assessment.
                          • David
                          • NJ
                          Regarding the immigration issue, Jan Brewer was in trouble before that legislation passed and she cruised to reelection. I think the anti-illegal immigration measures are very popular with voters across the country and the support of politicians like Gov. Brewer could turn out to be very beneficial for Mitt Romney who may be able to use the immigration issue to connect with blue collar voters.
                            • sirius
                            • Minneapolis
                            She won by 12 points in a strong Republican year.
                            PPP showed her approval split in February (46/47). That's about where the supposedly 'hated' Obama was at the time.

                            It's not 2010 anymore. 2012 isn't a mid-term election, either.
                          • Eugene Gorrin
                          • Union, NJ
                          Nate, you're right that the recent Arizona polls have this a tight race there - Romney ahead by 2 points in one poll, President Obama ahead by 2 points in the other poll. And President Obama's campaign has said it will be targeting Arizona in the fall.

                          But I don't think Arizona is realistically up for grabs this year. I think it will safely be in the Romney column.

                          However, due to demographic trends, unless the Republican/Tea/Evangelical party changes/softens its position on a whole variety of issues favored by the Hispanic population, perhaps in 2016 Arizona will be a "swing state.." More likely in 2020 Arizona will be a "blue state."
                            • rick d785
                            • Oregon
                            Hoe do people feel about Hillary Clinton?
                          • Maria Lopez
                          • California
                          Nate, your discussion about swing states & tipping points is interesting. Thanks. One other point I would make is that the Obama campaign presumably has multiple reasons for wanting to make Arizona a 'swing state'. By focusing on AZ, it forces Romney to do the same and takes his time away from other states where he needs to be. It can also give a psychological boost to the Democrats that they are competitive there. It also builds a foundation down the road for future Democratic nominees even if Obama isn't successful this time.
                            • steve vaillancourt
                            • manchester, nh
                            The map you show is a perfect example of why New Hampshire could well be the swing state even though we have only four electoral votes. Take Arizona away from Obama on your map and he's down to 268. Add NH back in for Obama and he's up to 272. Voila! Your perfect small swing state.
                              • Tim Page
                              • California
                              It was interesting to read your novel definition of "swing state." Me, I count the number of swingsets in a state before I use the term. Hey, why not, since we're playing around with the English language already?
                                • Edward
                                • NY
                                Do you have a better definition of swing state? Because Nate's makes a lot of sense.
                              • ibLoG
                              • Canada
                              If Hispanics turn Arizona into a blue state this November, GOP will not second guess about the Dream Act.
                                • Anthony N
                                • NY
                                The 2012 electoral college map at the end of this piece demonstrates the dilema facing Gov. Romney. Barring a massive GOP landslide, he has only one route to victory. He must "win back" Va., NC, Ind., Ohio, and Fla., and hold on to Ariz. or Mo. The President has far more possible winning combos to get to 270.
                                  • Mark
                                  • Tucson, AZ
                                  I have lived in Arizona since 1983. I know that I live in the only Democratic stronghold in Arizona, Tucson. However, I have yet to see one Romney bumper sticker yet. I see plenty of Obama bumper stickers! Obama 2012!
                                    • The Artist Formerly
                                    • The Blogosphere
                                    This reminds me of the sales guy who told me that McCain was going to win his state because he saw plenty of McCain signs and no Obama signs.
                                    McCain did not win his state, which was Massachusetts.

                                    The point: though I hope you're right, we can't place too much faith in your small-sample observations.
                                    • Hayford Peirce
                                    • tucson, az.
                                    I too live in Tucson and I certainly haven't seen a single Romney sticker -- and I don't think there have been many Romney signs anywhere. But there are a ton of Ron Paul signs all over the place....
                                    • Bob Geiger
                                    • Silver Spring, Maryland
                                    I think this speaks to something about the Romney campaign, it's lack of grassroots organization and grassroots enthusiasm. I drove across New Hampshire two weeks before the New Hampshire primary and saw only one yard sign for Romney. He won New Hampshire in the primary, of course, but the lack of signs was still striking, and I think telling.

                                No comments:

                                Post a Comment