Perfect Storm II?

I now divert from the realm of politics for a moment to alert the readership here that I may soon be dead.

Kidding, of course.  But – and this forecast is coming to you from the land of ridiculous model uncertainty and high stakes probabilistic forecasting – there exists an increasing potential for one of the worst east coast cyclones in recorded US history.  I strongly emphasize the word POTENTIAL, here, because the various models we use to forecast weather conditions 3-6 days in advance are all over the place, and the source of the diverse model solutions is not easily pinned down to one key forecast element to watch.

I’ve been monitoring this situation since last Friday when the long range models began showing the potential for both a major tropical storm threat and a rapid trough amplification over the Eastern US.  So let me lay it out for you.

In the Caribbean and just South of Cuba, we have a rapidly intensifying Hurricane Sandy-  it’s tropical envelop of circulation is extremely large and it is sitting over some anomalously warm waters (the Caribbean has been largely untouched by tropical troubles this season due to a weak El Nino that helped cut down tropical waves in that region).  It is expected to continue to deepen for a time over the Caribbean and eventually the Bahamas as it makes a strafing run along the Carolina coast.  And that is where the troubles begin.

Because as this storm lifts N or NNE, an intense ridge of high pressure is forecast to blossom south of Greenland associated with a blocked flow regime over this region.  Aficionados of winter weather will know that blocked flow over Greenland frequently forces storm systems dropping down from W Canada to amplify into deep troughs over the NE US.  And that is just what the models now agree is likely to happen.  As the flow becomes blocked, a shortwave trough is expected to approach from the west and then explode into a very deep trough over the Eastern US.  If the timing is right, this rapid deepening combined with the blocking ridge over Greenland may act to steer Hurricane Sandy NW or NNW instead of NE and out to sea.  And as Sandy begins undergoing a transition from tropical cyclone to nor’easter, if the timing is just right, it may explosively deepen (conditions in the upper atmosphere could come to favor natural cyclonic turning and upward motion).

I will begin monitoring this twice daily here, because this has the potential to be truly catastrophic.  I would place the odds at a coastal impact at 30-35% and the catastrophic scenario at at least 10%.  The NWS is now releasing extra upper air soundings until the event comes into better focus – that is an almost unprecedented event and it tells you how deeply concerned they are.  You hear folks talk about how bad it would be for a category 3 hurricane to hit NYC?  This actually has HIGHER upside for damage to NYC because the winds would last longer and mix better to the surface in a cyclone as opposed to a hurricane, whose winds tend not to mix to the ground well after landfall.

Just be on your guard…because this blog could end up a man down for a long time if this hits the way some of the models are now suggesting.

Advertisements

Swinging Long Shots – RIGHT: Arizona, Missouri and North Carolina

It says something about polling methods and the choices people make regarding the reporting of those polling methods that the entire blogosphere (virtually) would agree with my long shot left for the left but that all three of the state that make my list of right-leaning long shot swingers would be viewed skeptically.  RealClearPolitics does a poll of the polls and presents state and national polling data comprehensively.  It’s very useful for picking out trends in the polling data, but I caution users not to real it too literally because, for a number of reasons, state poll averages have a tendency to be unreliable and most polls have a democrat bias due to the challenges of collecting a proper sampling of the likely-voting electorate.

Until a recent surge by Romney following a strong debate performance on October 3rd of this year, every one of these three states – North Carolina, Missouri, and Arizona – was considered potentially in play for Obama.  After Mitt made his proper introduction to the American people to counteract the negative advertising, things seem to have returned to a more typical status quo entering a hotly contested election.  But let’s tackle these states quickly here to show what it is about the population that makes them unlikely to flip without a huge nationwide or demographically driven political change.

ARIZONA – County by County

Representative County: Navajo (Largely unpopulated) – here’s the kicker though…Maricopa (Phoenix) is actually nearly as representative…with a LOW bias in slant (in other words, the biggest city in Arizona votes more conservatively than the state as a whole)

Who lives in Phoenix – let’s look at the demographics:

%WHITE: 77.4 (average US major city: 62.6)
%POOR: 11.4 (average US major city: 23.5)
%RETIRED: 20.5 (average US major city: 16.6 – retired folks rarely live in cities other than Florida)
%MILITARY: 2.5 (average US major city: 1.1)
%MARRIED: 58.6 (average US major city: 40.9)

Phoenix is not your typical big city.  This town is older, more married, happier, wealthier, and less diverse than normal.  And even the Hispanics of Phoenix vote more toward the political center than average.

It just does not play well for someone like Obama who relies heavily on client-specific interest-group politics.

If you can’t sway Phoenix and its’ roughly 60% of the state vote even to the CENTER..you can’t win Arizona.

MISSOURI – County by County

Representative County: Clay (Kansas City suburbs) – though this county has a roughly 3-point high bias, it manages to achieve this high bias largely in the Gore/Bush election of 2000 – otherwise it predicts electoral outcomes well.  It should be regarded as conservative-tilting, rather than a pure swing county.

We’ve seen other swing states that have this look – the population centers lean left, the county is hard right – but Missouri’s suburban counties largely swing or lean right, rather than leaning left.  Take a look at the slant grouping data by cycle:

1 25.90% 25.54% 25.35% 36.05% 31.80% 34.25%
2 22.80% 23.52% 24.02% 43.36% 40.34% 43.48%
3 13.54% 13.90% 14.27% 50.59% 48.36% 51.00%
4 20.94% 20.18% 19.38% 52.70% 54.63% 60.00%
5 16.81% 16.86% 16.98% 66.54% 65.44% 69.71%

The conservative parts of the state were unmoved by Obama’s national surge – if Obama couldn’t move them…I’m not sure who could.  Those conservative parts of the state also push as far left as they can in most cycles within their respective slant groupings…so this isn’t just a large, lightly populated conservative leaning region…this is a VERY conservative area.  A very conservative area that makes up 48%-49% of the state’s vote totals.  So in half the state, you start out assuming you’ll give up 30 points…which means in the other half, you’d have to get back 30 points.  But the other half includes swing counties that often lean RIGHT, not left…and only 17% of the popular vote in MO comes from areas that lean hard enough to the left to balance out conservative regions.  In short – this state almost supported Obama…and he and McCaskill undertook a HERCULEAN effort to get out the vote here.  That liberal wave election of 2008 was a historic event…one that is unlikely to be repeated at any point in the near future.

And finally…there is North Carolina.

One part Appalachian, three parts Southern, one part academic (the research triangle).Barack Obama won this state by getting an unprecedented turnout in the research triangle among youths and an evne more unusual African American turnout (26% of the state’s population, 21% of the popular vote – normal here has been 15-16%).  But when you get right down to it…NC is a conservative (typically southern) state on the whole.

Here is the slant group tabular data:

1 33.00% 32.13% 31.69% 33.37% 32.01% 36.36%
2 15.71% 15.47% 15.18% 42.30% 40.95% 45.72%
3 24.04% 24.70% 24.91% 46.66% 48.15% 55.34%
4 20.41% 20.63% 21.17% 50.37% 51.28% 60.45%
5 6.84% 7.07% 7.05% 63.95% 66.04% 72.42%

There was a huge 10 point shift for Obama in slant group 4 (areas that have been swing counties in the past that suddenly voted like liberal bastions).  Those kinds of shifts only sustain themselves if there was a wholesale change in the demographics of the region – which in this case there was not.  The state is growing, but the proportions of the various demographic groups haven’t changed much.  And once again, there is a near controlling interest in the state’s voting stock held by hard right and right-leaning regions.  I’ll spare you the map, as this post is already of great length, but watch the Wilmington area – the county containing this coastal resort city (New Hanover) has represented the state well, and with very little slant bias.  Polls show Romney running away with NC now and Obama has abandoned the state to concentrate on more attainable targets – for what that’s worth.

Swinging Long shots: LEFT Oregon and Michigan

I’m not going to spend as long on the long shots, but I think the readers ought to have a quick look at the fringe states that the media is completely ignoring or erroneously fixing on throughout this campaign season.

On the left, there are two long shots – Oregon and Michigan – both of which have ingrained population distributions and political behaviors that haven’t much changed over the last three cycles and that likely remove them as Romney targets.  On the right – you’ve got three long shot states – Missouri, Arizona and North Carolina.  We’ll look at the left-leaners today and the right-leaners tomorrow.  And we’re going to just throw the data out there and then summarize the main problem for the trailing candidate.  Then we’ll give the formula you’d need to see to have the long shots flip.

Oregon – County Map (2000-2008)

Representative County: Columbia (Portland suburbs)

Oregon by the numbers:

1 12.78% 12.09% 11.63% 30.79% 31.68% 36.51%
2 19.48% 19.61% 19.95% 41.56% 42.69% 48.59%
3 19.47% 19.09% 18.69% 48.25% 47.76% 53.73%
4 26.43% 26.98% 27.21% 53.52% 55.57% 62.17%
5 21.83% 22.24% 22.53% 67.63% 70.99% 77.39%

Hopefully you can see the problem with this state.  Portland, Salem and Eugene dominate the state’s population – have been gradually trending left for the last 20 years, and were enough to give Gore the state (by a fraction of a point) 12 years ago.  Obama’s big score here is probably just a bump…but the drift is real, and you’re not going to enter the City of Roses and reverse the liberalization of that place now.  In order to win Oregon, you need to replicate the 2000 formula and then you need to get the population surge in Portland to reverse at the polls (which is unlikely), and then you need to increase voter turnout in the sticks…which is just as unlikely.  I think the state is lost despite the polls showing it as tantalizingly close.

On to Michigan:

County Map

Representative County: Gogebic (adjacent to liberal Duluth, MN)

This looks playable, right?  You think – well most of the state isn’t on the fringe left and there are more right-leaning regions than swing regions – this should work, right?  Wrong.

1 7.01% 10.73% 7.36% 34.98% 34.46% 42.66%
2 19.39% 18.82% 19.37% 41.20% 41.15% 48.99%
3 35.71% 34.33% 35.44% 50.00% 48.58% 54.84%
4 8.94% 8.57% 9.03% 53.81% 53.59% 60.28%
5 28.94% 27.55% 28.80% 67.43% 66.66% 72.38%

Two problems: the conservative vote is too soft – most of the conservative parts of the state just lean right…democrats pick up a good 40% of the vote outside of the cities.  And actually, the populous parts of the state are in either liberal Detroit/Lansing or in swing country…a republican needs to achieve a large cultural swing, not just pick at areas that are vulnerable like in PA or VA – he needs to play to all types of people, and that includes unions and manufacturing sector employees that may not lean HARD left, but consistently lean left.  The coalition in this state must be broad…if Romney wins here, it will prove that he has indeed appealed to the entire nation.  These days, that is awfully tough.

Swingers: Florida and New Hampshire

We’re going to tackle the last two swing states in the core eleven in one post, because New Hampshire is tiny and does not take much time to analyze and the election now approaches rapidly, so I’d like to gt to the outer five and then to the publication of the county-level statistics state by state – things like the population slant adjusted metro area ratings and the county by county bias and vote percentage stats so you can follow along on election night and make some back of the envelop projections as the data comes in.

First, let’s tackle New Hampshire – and I don’t want to spend a ton of time on this most conservative of New England states – its 4 EVs and 10 counties are, according to the ensemble of polling data, breaking for Romney by a few tenths of a percentage point at the moment, but political insiders still county it as a relatively likely Obama win based largely on how Obama did here in 2008.  I believe they are mistaken.

Here’s the map:

Cities of New Hampshire:

  1. Concord (11% of vote – leans left, but barely)
  2. Manchester (leans slightly to the right)
  3. Nashua (right down the center – Manchester and Nashua account for ~29% of vote)
  4. Portsmouth (right down the center – ~23.5% of vote)
Now the three-cycle average slant groupings by county (2000-2008)
The base color is the official slant grouping, but for this state, I’ve done a bit of stippling to illustrate the reality that the vast majority of the state’s population resides in counties which are, for all intents and purposes, swing counties or darned close.  The liberal influence of Vermont (the state that scares the hell out of New Hampshire) does bleed across the state line in the western (and lightly populated) rural counties, but the meat of the state has been keenly politically aware and painfully undecided for many cycles.
Democrats have an edge in party registration here, but that’s been true for a long time and yet here we are, with three-cycle averages that include an Obama bounce and are nonetheless very close to ideological parity.  And the key is that the state does not have any traditional urban centers – the places where liberals go to rule like kings, never questioned by a populace cowed by urban political shenanigans (ask Tony Rezko about thow that works – or you could just ask his wingman Barack Obama).  The population of New Hampshire lives in a world where all voices are usually heard.  The democrats don’t have safe havens that eat 40% of the vote the way they do in most other swing states.  And in elections like this one where the incumbent is wildly unpopular (by the standards of the typical incumbent President – whose favorability is usually above 55%) and has actively campaigned on a platform of negative attack ads and clientalism – and in a state where clientalism doesn’t apply (New Hampshire is not particularly laden with affronted minority groups and there are far more married women – who break for Romney – than there are single women – who break for Obama), I’d say this is one of Romney’s EASIEST pick-up opportunities outside the already pocketed North Carolina and Indiana.  This is especially apparent when you look at how Obama built his 10 point majority in 2008:
(slant grouping, vote percentages in 2000 – 2008, and slant in 2000-2008)
2 4.71% 4.77% 4.73% 42.01% 44.00% 50.59%
3 59.47% 58.93% 58.78% 48.33% 48.35% 51.73%
4 29.96% 30.20% 30.47% 51.25% 54.49% 59.80%
5 5.87% 6.10% 6.02% 55.76% 59.75% 63.95%
All of these group slant figures are WAY lower than normal – most states return group five slants over 65% because most states have at least one rotting corpse of a city that turns out a million zombies to vote democrat and then return to living in misery.  Before 2008, the state’s lone group five county was a group four county.  Most states return group four averages near the 57% mid-point for that bracket because those group four counties are a mix of inner ring suburbs (filled with rich white liberals) and small cities (filled with bureaucrats who subsist on government funding).  Most of New Hampshire’s group four counties were swing counties before 2008.  And most states turn out swing counties that sit near the 50/50 line, but in New Hampshire, the rightward slant is noticeable – that is, except in 2008.  The representative election for New Hampshire is probably 2004 – Kerry carried it by 1.4 points, largely due to a lack of GOP enthusiasm for Bush here…but it was a bitter fight right up until the end.  It’s naive, IMHO, to assume that the three-cycle leftward drift is permanent when the demographics don’t support that conclusion and when the state gave Obama a modest bounce compared with other swing states.
The representative county here is Merrimack (home of Concord – a left-leaner that tracks election results well but with a high bias of 1.7 points)…that’ll do for me.  Watch Merrimack county and subtract one or two points and see how it goes.
Now…on to Florida.  This state imports old people and tourists and exports election night nail biting and hanging chads.  They probably have some other revenue, but at the moment, I don’t much care. 🙂  Three things you need to know about Florida’s population:
  • The elderly make up a solid 18% of the state’s population – higher than any state in the union (Arizona is now a close second, FWIW) – they tend to be registered Democrats from New York…but they tend to vote like undecided swing voters, leaning further right with age.
  • Hispanics make up roughly 30% (!) of the state’s voting aged population, but a large number of them are Cuban and Cuban voters do not nearly as far left as do most Hispanics (Castro is among the most hated people of all time in Florida and that does not play well for Democrats).
  • Florida has very high population density, but votes very conservatively compared to normal for high density regions – there are a number of reasons this might be (Democrat corruption in the war on drug cartels in Miami, the presence of anti-Castro anti-socialist voices like Marco Rubio, the high voter turnout among the elderly, etc), but the numbers don’t lie and haven’t changed much over the last several cycles.
Population Centers:
  1. Pensicola (~3% of vote)
  2. Tallahassee (~1.5-2% of vote)
  3. Jacksonville (~4.5-5% of vote and rising)
  4. Orlando (~7-8% of vote and rapidly rising)
  5. Tampa/St. Petersburg (~11.5-12.5% of vote and rapidly declining)
  6. Ft. Myers (~3% of vote)
  7. West Palm Beach (~7% and slowly dropping)
  8. Miami/Ft. Lauderdale/Miami Beach (~18.5-20% and very rapidly declining)
Here’s the 3-cycle map for the state with population centers marked:
As I noted before, Miami-Dade County – the most populous in the state – is not like many other urban centers with a dominant democratic machinery – it’s the suburbs (Ft. Lauderdale and West Palm Beach) that really tip the scales, and did so en masse for Obama.  On the other hand, the state capital in Tallahassee is solidly democrat controlled – the main minority population there is African American as opposed to Cuban, but and that makes a big difference in the political slant – it’s more of a traditional Southern city in that regard (and note that the panhandle votes much more like the south than does the peninsula).
The state slant was neutral in 2000 and +5 for Republicans in 2004…so how did it go so far back toward the Democrats in 2008 (they won it by 3 in Obama’s sweep)?  Let’s look at the trend diagnostics and then try to figure what, in the demographic data, would support such trends:
Here is the complete list of counties that gave Obama more of a bump than the national average (which moved from R+2 to D+7 (9 point swing):
  • Escambia
  • Duval
  • Clay
  • Pinellas
  • Orange
  • Osceola
  • Polk
  • Hillsborough
  • Hardee
  • Lee
  • Hendry
  • Miami-Dade
In total, the state moved left about as much as the nation did, and what’s remarkable is that Obama’s gains were clustered in the cities almost exclusively, except in Hardee, Clay, Polk and Osceola counties where clusters of Hispanic or African America populations above the state norms (outside of Miami-Dade, of course) alter the demographic picture  Obama had a very clear strategy in Florida in 2008 – he went after young Hispanics and he attacked the cities, ignoring the rest of the state, and it worked.  At a time when most of the south outside of African-American-laden districts actually tilted further RIGHT…Florida tilted left.  This year, he’s trying to scare the older voters with threats to their medicare…a strategy that has not worked at all and has cost him time that he could have spent wooing Hispanics again…and now polling is showing his support among Hispanics and the youth of the state corroding badly.  On to top of that:
1 12.45% 12.92% 13.04% 33.16% 29.83% 32.80%
2 26.58% 27.42% 27.33% 43.80% 41.37% 45.75%
3 24.85% 23.31% 23.58% 51.26% 46.55% 52.97%
4 18.96% 19.45% 20.01% 53.04% 52.23% 58.59%
5 17.16% 16.91% 16.04% 66.53% 63.23% 64.88%
Obama’s big gains in the state were largely in swing and left-leaning counties like Miami-Dade, and Pinellas/Hillsborough…he didn’t get much of a bounce elsewhere in the state and the bounce he got in those swing counties is probably not sustainable, given the long term trend toward dixiecrats voting Republican and seniors voting more conservatively with time.
The bottom line: conservatives match liberals here person for person and they vote more conservatively than the liberals vote liberally, conservatives were unmoved by Obama and swing voters will be unmoved this time around.  This state is a mortal LOCK to go for Romney…that’s just how I see it.

Notice to Those Nitpicking the 9/12 Rose Garden Speech: Focus on Obama’s ACTIONS, Not His Words

It doesn’t matter one whit what Obama said the day after the attack on our consulate in Benghazi. It matters what he and his administration did in the weeks afterwards. Instead of taking the matter seriously, Obama jetted off to Vegas for a campaign event. Instead of acknowledging that the attack was an entirely premeditated, Al Qaeda-run operation, Obama and his underlings repeatedly insisted that a stupid You Tube video was at fault. One speech cannot erase the absolute clusterfrack that was the White House’s response to the aforementioned act of war, and don’t let the left convince you otherwise.

Co-author’s edit:

The above is completely true, but you don’t have to ignore the Rose Garden speech to come to the right conclusion.  In that Rose Garden speech, Obama excoriated the YouTube video, said they were still gathering information on the Libyan attack, and apologized, saying the government had nothing to do with the video and believed in respecting all religious faiths and said NOTHING reflecting any sort of certainty regarding the Libyan attack as SPECIFICALLY a terrorist attack.  The “acts of terror” phrase was meant “in general” as a warning not to engage in further violence over the YouTube video…clearly he was not calling out Al Qaeda for the Libyan attack or he wouldn’t have been talking about some damned YT video.  To maintain otherwise is to be deliberately STUPID.

Original author’s edit:

Actually, the Rose Garden speech didn’t mention the YT video as far as I know. My co-author may be confusing that speech with the one he delivered at the UN. But yes — context matters. Obviously, when Obama used the phrase “act of terror” on 9/12, he didn’t actually mean that the attack in Benghazi was a terrorist attack. He was using words that sounded “tough” and “reassuring,” but if he actually believed those words, he would’ve behaved accordingly.

Swingers: Virginia

We’ve gone through most of the swing states we planned on covering from the “true swing” category – all that remains are Virginia, Florida and Hew Hampshire.

So let’s get to it with the extremely complex state in which the RightFans clan makes its home – Virginia.  The home of Presidents best known for its modern day population diversity, its heralded post-secondary education system, its deeply divided state legislature, and its role in the American Civil War as the beachhead of the Confederacy – a legacy that in some corners still lingers.  It’s the birthplace of General Lee, Thomas Jefferson and George Washington, among other big figures in American history, and it has always been at the crossroads of competing cultures – once North and South, now elite and rural (not to mention the presence of a strong military contingent, a large and active Asian immigrant bloc and the rise of Hispanic immigration to give it even more color).

Remember how Ohio achieved barometer status by being the Midwest in a tight, heavily populated package?  Virginia has a similar patchwork demographic pattern – it’s an amalgam between Appalachia, the deep south, and the Northeast.  I’ll mark out those regions when we take a look at the maps.

Populated Regions:

  1. Norfolk/Virginia Beach/Newport News (~10% of vote – leans left slightly overall)
  2. Richmond (~2.5% of vote – very liberal)
  3. Fairfax/Arlington/Alexandria (~19-20% of vote and rapidly declining – solidly liberal)
  4. Roanoke/Lynchburg (~3-3.5% of vote – leans right)
  5. Charlottesville (~2% of vote – solidly liberal)
Representative County:
Montgomery (home of Blacksburg and Virginia Tech) – this may be a fluke.  The county happened to experience a surge in youth voting from the college at the same time the state as a whole tipped way left due to increased voter turnout in the urban areas and government sector (Northern VA suburbs of DC) – two things that aren’t completely unrelated but probably cannot be counted on to perfectly correlate going forward.  Here’s a list of runner up counties/cities that crack the top twelve in terms of mean absolute error:
  1. Montgomery County (1.26) – SW VA
  2. Essex County (1.38) – SE VA
  3. City of Hopewell (1.80) – Tidewater VA
  4. City of Radford (1.91) – SW VA
  5. Buckingham County (2.13) – Central VA
  6. King and Queen  County (2.18) – Tidewater VA
  7. Henrico County (2.19) – Richmond Suburbs
  8. City of Winchester (2.20) – Shenandoah Valley
  9. City of Manassas (2.35) – Northern VA
  10. Loudon County (2.46) – Northern VA
  11. Chesapeake County (2.49) – SE VA
  12. Prince William County (2.83) – Northern VA
All regions are well represented, and each region is trending the way it does for different reasons.  In the Tidewater area, the leftward lurch of 2008 in otherwise conservative-leaning areas can be explained by the combination of increased African American turnout and increased turnout in liberal college campuses among youths (VA has many college towns in the Tidewater region).  In the Richmond and Norfolk suburbs, African American turnout spiked as it did through the Southern states.  In Northern VA, the trend is leftward generally, but it really stands out in the expanding suburbs which used to vote Republican and, as suburban sprawl takes Fairfax County voters outward, now vote further left.  And in the rural Western parts of the state, the few counties that tipped leftward did so either due to increased youth turnout or increased turnout by union members who had been voting more conservatively but were still registered democrats.
The picture, however, is pretty darned complicated – and here it is:
Three cycle average county slant: 2000-2008
One thing to point out with this map – it’s a little deceptive because Virginia does something no other state in the union does – it reports its incorporated towns and cities separately when it tallies the final vote totals by county.  There are a few major cities in other parts of the county that get reported separately (St. Louis City vs. County, Baltimore City vs. County, etc), but in VA, there are 39 incorporated towns and cities that report separately – often the lone population cluster in a rural county will report separately from the sticks.  So, for example, in Appalachia, whereas in WV, the voting pattern is largely centered on slant group 2 (conservative leaners), here it looks mostly very conservative (group 1), but the tiny dots you see peppering the state vote to the left of the counties in which they reside most of the time and if, as is done in WV or OH or PA, you report the county votes INCLUDING the incorporated cities, you’ll get a lot more pink than red.
Let’s have a look at the trend analysis:
Appalachia generally continued a gradual slide toward an increasingly conservative vote, especially in the panhandle while Northern Virginia drifted further left than the rest of the state (except Charlottesville – a town benefiting from more Federal funding than usual due to an active research industry), but it is important to point out that the people fleeing Fairfax county are not, in fact, as liberal as the people staying put, since Fairfax slammed HARD to the left (by almost 10 points more than the nation as a whole did) while the other suburban counties gently nudged left, implying that despite the sprawl I mentioned above, the outer ring suburbs of DC are not yet lost to Republicans.  In fact, the people on the ground in Virginia have been telling me that Romney enthusiasm even in increasingly liberal places like Prince William and Loudon Counties is quite high and more visible than current Obama support, for what that’s worth.
Let’s chart out the state’s slant group and population drift over the last three cycles for your benefit as well (remember, this chart lists three-cycle slant groups, voting percentages in 2000, 2004 and 2008, and then voting slant in the group in 2000, 2004 and 2008):
1 18.36% 18.59% 17.89% 34.12% 31.52% 35.63%
2 27.97% 28.38% 28.16% 41.72% 40.02% 46.83%
3 17.42% 18.02% 19.08% 44.82% 46.85% 55.28%
4 22.09% 21.27% 20.68% 50.35% 53.36% 60.98%
5 14.15% 13.74% 14.18% 64.23% 65.98% 73.68%
The conservative parts of VA generally held the line and remained conservative, not giving a large bounce to Obama, but swing and liberal counties leaned far to the left, giving Obama a 7-9 point bounce in over half of the state’s electorate.  This, combined with the population move away from Fairfax and into Prince William, Loudon, and Fauquier counties (accounting for the entire shift from left-leaners to swing counties) makes the swing counties even more critical now than they were before Obama’s Presidency.  So keep your eye on the outer ring suburbs of DC, Richmond, and Norfolk on election night…you’ll learn a lot doing that.

Swingers: Pennsylvania

Despite being spammed by some swingers website using the term in its least attractive light, I’ll stick with the horse that brung me as the saying goes. 🙂

Today, we move on to the Keystone State – a state that in urban and rural towns alike looks war torn, a casualty of the death of American manufacturing and the victim of liberal political machinery in the cities and the total domination of the EPA over energy policy.  Here’s the perfect metaphor for you.  When you drive from DC to Philly (as my family so often does to see relatives in the NW Philly suburbs), you encounter a moment where the roads go from pristine to precarious.  At that moment, if you look to your right at the roadside signs, you’ll see one bidding you welcome to Pennsylvania.  The failure of PA state agencies to maintain infrastructure (roads, bridges, tolls and stable power) or provide adequate services like winter snow removal and sanitation echoes the failure of the Federal Government to provide adequate resources to the military or exploit the nation’s energy resources to keep us off of oil from the Middle East.  It’s not caused purely by the left’s obstructionism and poor budget-wrangling skills – if it were, than the roads in Maryland would be just as bad as they are in PA.  It’s caused by an unhealthy polarization – a stalemate between conservatives in the “T” (the rural counties at the heart of the state) and liberals in Philly and Pittsburgh over what the priorities of the state must be when setting a budget.  It’s a culture war that’s playing out all across this nation, but viewed in microcosm here, it looks particularly stark.  In the state legislature, there is constant division between two groups of people that simple do not understand the other side and make no effort to do so because their lives are too different.  Obama himself demonstrated the left’s refusal to really understand the people of the T when he famously accused them of clinging angrily to their guns, their religion, and their racism to deal with hard times.

So let’s take a look at the state over the last three election cycles and expand upon the divided (and difficult to change) landscape a bit further.  There is one crucial trend that may change the way the state plays in future elections, and we’ll look at that in a moment:

Population Centers:

  1. Erie (~2% of vote and dropping)
  2. Pittsburgh (~11-12% of vote and rapidly dropping)
  3. Altoona (~2% between Blair and Cambria counties)
  4. Harrisburg (~4% of state vote between Dauphin and Cumberland counties)
  5. Reading (~3% of vote and slowly rising)
  6. Scranton/Wilks Barre (~4-4.5% of vote and dropping between Lackawanna and Luzerne counties)
  7. Allentown (~2% of vote)
  8. Philadelphia (11.5-12% of vote and slowly rising)
Representative County: Northampton (Allentown)
Users should note that we rated this county in slant grouping #4, but it BARELY crosses the 53% threshold and in previous election cycles has been much more of a swing county – which does perfectly represent the history of PA as moving from genuine swing state to left-leaner in recent years.
Liberals haven’t really tried to attack the state’s rural counties outside of college towns like State College (We Are…Penn State!) and crumbling factory towns like Altoona, Erie and Scranton in the last several Presidential elections.  But Republicans have opened up a new front in the war for PA’s 20 electoral votes, as demostrated by this trend map from the Election Atlas:
These trends are relative to the national average trend (if the national electorate moved 5 points left, and your county moved 1 point left, you’d get an R+4 trend).  And once again, this guy’s site is bass-ackwards, and shows liberal trends in hot colors and conservative trends in cool colors.  So what you’re seeing is the end of the era of the T.  As Pittsburgh collapses and its union presence travels elsewhere to blight another city with hopelessly expensive labor contracts and political corruption, the entire Western part of the state is trending hard right.  The suburbs of Pittsburgh, outside Alleghany County, moved 2-3 points left between 2000 and 2004, and another 3-5 points left between 2004 and 2008.  The same trend is occurring in the Eastern counties of Ohio – areas that used to swing more and are now voting like West Virginia (which has not chosen a Democrat President since Clinton and is polling further right this cycle).  In fact, our ground spotters in PA have been telling us all season long that the western part of the state is showing far more enthusiasm for Romney than they showed even for McCain, and the state polling agencies (as opposed ot the national outlets which bias a bit too heavily toward urban areas) are showing Romney holding much bigger leads in the Pittsburgh suburbs.
At the same time, the wealth of Philly is expanding west and filling in formerly rural parts of SE PA, and, as such, the upper-middle-class liberal vote share is increasing there, so the state is still not going to be easy to win.  But unlike a number of the other states I’ve looked at, the battle lines do appear to be moving here and it remains to be seen which changes will matter most in the long run.  If you take the state in aggregate, here’s what the numbers (by 3-cycle county slant group) look like:

1 20.55% 21.02% 20.92% 34.88% 33.80% 39.87%
2 8.98% 8.73% 8.37% 44.62% 42.49% 44.18%
3 23.18% 23.46% 24.06% 49.36% 49.00% 53.25%
4 33.90% 33.27% 32.91% 55.93% 55.51% 58.42%
5 13.38% 13.53% 13.74% 78.83% 77.48% 80.89%

Population-wise, there’s been some leakage out of left-leaning urban districts like Wilks-Barre, Scranton, Erie and Pittsburgh and some move toward swing towns like State College, Harrisburg and Allentown, with no penalty to the state’s slant behavior by group other than a modest Obama surge (less so here than other state).  It’s true, in fact, that Conservatives reacted negatively to Obama’s comments about them and rejected his national surge in enthusiasm, as did the Pittsburgh suburbs that are traditionally swing counties.  It was only in Philly’s swing counties where Obama got any kind of bounce.  The state does not seem poised to enthusiastically support Obama…if the challenger proves himself worthy, you could see things flip here.

For what it’s worth, my family out in Montgomery and Bucks counties has been claiming that the NW Philly suburbs are trending toward Romney  – much more street-sign presence for the GOP than for Obama.  To win the election, we’ll need to hold places like Bucks County, the Allentown area, and the Pittsburgh suburbs, and the youth vote in the college towns will need to drop.  Centre County leans right most of the time, for example, and leaned left for Obama due to increased youth turnout from Penn State.

With the landscape shifting, however, I actually think there is a better chance Republicans make progress here than in many of the other swing states we’ve looked at this far, with the possible exception of Colorado and Ohio.

Swingers: Ohio

It seems like, every cycle, this is the state that decides the election.  In fact, no one has won the white house without carrying Ohio since before FDR came to power, other than one miss – the 1960 election of John F. Kennedy (!).  This is not likely because the state’s 18 electoral votes are any more decisive than Pennsylvania’s 20 or Florida’s 29.  This is because Ohio is, in one state, the entire Upper Midwest.    To the east, there are aging, rusted out steel mills and an assortment of manufacturing locations not dissimilar to what goes on in Pittsburgh, PA and Wheeling, WV  To the south, there are bluegrass festivals and rolling hills filled with small towns not distinct from the towns of Kentucky.  In the west, apart from Cincinnati – the state’s most conservative large city – there are farms and even livestock, like you’d find in Indiana, and in the north, there are Toledo and Cleveland, which bear a striking resemblance to the crumbling crater that was once Detroit, with similarly ugly crime statistics, annual deficits and economic stagnation.  So, if you squeezed the Midwest up like a paper ball, you’d get Ohio – and you can’t win the election if you can’t make a dent in the Midwest, since every state in the region is heavily populated and packs double digit electoral votes.

So here we are – with positive momentum from the rumble and the crumble that was October 3rd’s presidential debate – and Romney’s probably flipped Florida, North Carolina, Virginia and Colorado.  Meaning he’s 13 electoral votes short of the oval office.  He can get those electors without Ohio, but here’s the question – do you suppose he can win Iowa Wisconsin if he’s unpopular in the ultimate Midwestern state (Ohio)?  So some will be saying that Ohio is not a must-win for Romney, but I would say that the nation’s electoral barometer will remain with its perfect batting average even so, because if Romney does not win Ohio, I don’t think he’ll win Wisconsin and Iowa…and the alternative would be for him to win an even tougher state like Pennslyvania or take down three small states that are tough to get (like New Mexico, Nevada and New Hampshire).  All of which is to say that either Romney is going to lose by a mere handful of electors…or he’s going to win by perhaps 50 and it will be called a blowout.

So let’s have a look at Ohio – how is the nation’s barometer looking over the last three cycles?

Population Centers:

  1. Youngstown (~2-2.5% of state vote and declining)
  2. Akron/Canton (~6% of state vote)
  3. Cleveland (~11.5-12% of state vote and declining rapidly)
  4. Toledo (~4% of state vote and declining)
  5. Columbus (~9-10% of state vote and rapidly rising)
  6. Springfield (~1% of state vote)
  7. Dayton (~5% of state vote)
  8. Cincinnati (~7.5-8% of state vote and declining)
Three-cycle county slant (2000-2008)
Note that the county best representing Ohio as a whole is Ottowa County – the southern suburbs of Toledo – directly at the intersection of Indiana life (and by extension the life of the Central Plains and Southern states balled into one amalgam) and the life of the corroding industrial belt that once powered places like Detroit and Cleveland.  Actually, the eastern half of Southern Ohio plays more like Appalachia, with a majority of registered Democrats, but with significant portions of that old democratic base now voting Republican in most elections, resulting in a slight, but pervasive and unchanging, conservative tilt.  So you can divide the state three ways – the Appalachian bordering piece that votes and looks much like WV, the rust belt counties that vote and look like Pittsburgh and Detroit, and the more pastoral chunk of the state that votes like Indiana.  But our representative county is resting right on the fault line between traditional Indiana living and the rust belt.  I would, however, urge you to look carefully at the conservative cities in Ohio – places like Cincinnati, Springfield and Dayton – to get a feel for how the state might break (Clark County, in particular – home of Springfield – very nearly took the top spot in the list of counties that predict state outcomes well).
Early voting suggests that Ohio’s population of registered democrats is down significantly from 2008 and that turnout in ultra-liberal Cuyahoga County is WAY off from the pace set by Obama’s ground game last time.  Which isn’t to say that Cleveland residents won’t vote…I’m sure the final numbers won’t be off by that much…but it does suggest a decline in enthusiasm among Ohio’s urban voters.
Here are the voting patterns of the last three election cycles by three-cycle-average county slant:
1 24.28% 24.90% 25.56% 34.38% 33.59% 37.65%
2 9.82% 9.66% 9.58% 41.99% 41.97% 45.56%
3 25.02% 24.37% 23.87% 47.56% 48.80% 52.14%
4 19.67% 20.36% 20.83% 52.78% 55.18% 58.98%
5 21.21% 20.71% 20.17% 63.46% 64.69% 67.48%

Ultra-liberal Cleveland, Youngstown and Toledo are losing population to places like Columbus, resulting in a reshuffling of the deck chairs.  Soon, Columbus may join Cleveland in bracket five.  If Ohio is a barometer for the state’s voting preferences, it may bode well for Republicans…the rust belt is collapsing and cities like Columbus are taking in the refugees, but in total, conservative counties have gained an extra percentage point of the state vote and liberal counties have lost a point, though the exchange comes at the cost of swing county votes being closer to the 50/50 line than they were before.  In general, ignoring Obama’s 4-point enthusiasm bounce and looking at the longer term trends, the state looks poised to remain hotly contested for the foreseeable future.  With Romney now within a point in most Ohio polls, this one will be worth watching closely on election night.

Swingers: Wisconsin

I was already a bit familiar with the voting patterns in the cheese belt because I watched the results pour in as Scott Walker attempted to survive a recall vote (and succeeded for the first time in history).  The state became a focal point in a battle over whether it was proper to limit the power and political influence of public sector unions after Walker and a team of Republican state legislators got elected promising to fix the state’s budget woes and did so by requiring state unions to pay more into their health care coverage and stripping them of their right to collectively bargain over anything other than base salary and pay automatic pay increases, as well as requiring public unions to end the practice of extracting mandatory dues and to give people a right to work a trade without belonging to that trade’s public union.  The package became known as a union-busting bill and for months, teachers unions spearheaded a massive protest in Madison that served as a blueprint for the Occupy Wall Street Movement that would later follow.  With the capital under siege and all but two democrats out of the state to prevent the quorum necessary to hold a vote on the bill, Walker and Wisconsin republicans were forced to wait it out.  When they finally passed it, they faced three separate legal challenges, all rejected, followed by recall elections in left-leaning state congressional districts and for Walker and his lieutenant Governor Kleefisch.  Walker survived the recall vote, but a subsequent state supreme court challenge has temporarily nullified the law while Walker’s legal team appeals to a Federal district court.

Amid all of this drama, there is a hard reality for democrats – Wisconsin – long a safe democrat state (interrupted by popular support for Reagan) – may be trending back to the right since 2008’s wave election.  Republicans survived the recall vote, for the most part, but it is not clear how much momentum Conservatives in the state will enjoy.  It is generally a good sign for a party when a number of its best and brightest minds are coming from a region – on the right, Wisconsin has recently been a breeding ground for some mega-talents that reveal a robust and spirited state-level GOP that is engaged and making gains.  Paul Ryan – lead voice of conservative intellectualism on matters of budgetary policy – comes from a left-leaning swing district that has comfortably supported him by margins no less than 64-31 since his first election.  Rience Preibus – head of the RNC – is a master fundraiser and spin-doctor, and is building a reputation as a leading party strategist as well – he too originates in Wisconsin.  Scott Walker is among the hottest rising stars of of the GOP governors association.  Rebecca Kleefisch will likely follow Walker as governor and is widely considered a leader among conservative women of the Midwest despite her relative youth.

Adding to the curiosity about whether Wisconsin might be playable for Romney in 2012, Romney selected native son Ryan as his running mate and almost immediately, polling drastically improved – now the consensus seems to be that Romney is trailing in Wisconsin by at most 2-3 points and that he’s running in the state almost uncontested as Obama is forced to save his resources to defend a crumbling firewall in Ohio, Virginia and Florida.  While polling in Wisconsin continues to improve in the final month of campaign season, Obama’s firewall is down to one.  VA and FL polls have trended strongly to the right as Romney has utterly buried Obama in spending on television ads and ramped up his ground game to match the formidable efforts of the incumbent in the White House.  But let’s familiarize ourselves with this crucial battleground and get the lay of the land.

Populous Regions:

  1. Eau Claire (~2% of vote)
  2. La Crosse (~2% of vote)
  3. Madison (~9-9.5% of vote and rapidly increasing)
  4. Green Bay (~4% of vote)
  5. Milwaukee (~16% and declining rapidly)
  6. Kenosha (~2.5% of vote and increasing)
  7. Fon De Luc (~1.5-2% of vote)
Wisconsin counties by slant grouping: averaged from 2000 to 2008
The keystone county this time is Iron County – on the far outskirts of the Wisconsin suburbs of Duluth, MN – it’s a quiet little county without a ton of votes to contribute, but it does a good job of predicting the behavior of both the rural conservative counties in the North of the state and the suburban and urban liberal counties in the South and West.  I also recommend watching the Green Bay vote tally and adding 2.5 to 3 points to the democrat vote to get a good feel for the state’s likely final outcome…Green Bay (Brown County) has almost perfectly predicted the precise state swing from cycle to cycle – it just votes a few points right of the mean for the state.
Also, notice a pattern yet in all of these states?  Well there are two things that are always true:
  • Urban areas tilt left of the state in general – very few cities defy this paradigm, and in fact, I found an obvious (and robust with 99.5%+ confidence) linear correlation between the natural log of population density and slant nationwide.
  • Urban backlash is the only thing that seems to break this nearly perfect correlation – even moderately populated suburbs near a very liberal city will tend to vote right of center and stay right of center even in elections like 2008, where the whole electorate is shifting.  I believe what’s going on here is that inner ring suburbs surrounding a major liberal city have a birds eye view of exactly what liberal policies get you in terms in stagnation, poverty, crime and education, not to mention taxation and job prospects…and they vote accordingly (for the guys who will fight against urban political corruption).  Obama would see this pattern and assume that the wealthy suburban elites support Republicans and deny funding to the city selfishly, but a look at Milwaukee suggests that the inner ring conservative bloc is actually populated with primarily middle class earners and not the rich.  So bear that in mind if you hear that argument.

Wisconsin counties by slant grouping: 2008 election
Note that the inner ring suburbs of Milwaukee held the line and voted for McCain in droves…the entire rest of the state moved left a good solid 8 points, shifting entire regions into the top liberal bracket and the vast open spaces of swing country into democrat leaners.  Yep…McCain took a brutal beating in Wisconsin.
You can see the shift even more clearly when you look at the state’s three-cycle slant grouped county statistics (once again 2000, 2004, 2008 election cycles in that order, first percent of state vote, then slant)
1 12.48% 12.30% 12.50% 32.35% 32.03% 37.09%
2 13.38% 13.76% 13.85% 42.36% 41.29% 48.78%
3 30.38% 30.46% 30.12% 47.52% 46.99% 54.82%
4 13.50% 13.60% 13.61% 53.71% 53.89% 61.17%
5 30.26% 29.87% 29.93% 62.01% 63.36% 69.61%
Liberals and learners shifted by a full 7-8 points, while conservatives shifted less and voted a little more, trying to keep it close.  I saw something similar in Nevada and New Mexico, and definitely spotted it in Minnesota near Minneapolis where regional tax-base sharing is clobbering conservative suburbs with huge tax increases.
But here are the facts, regarding the state since 2008:
In 2010, even the SW part of the state managed to send republicans to the state legislature (albeit only one or two).  Walker received a vote share that looked more like Bush (2004) – here, let’s just look at the map from that recall election (I won’t use slant grouping colors here…I’ll show you the image from an election data website I’ve been using for these analyses (you can find it Deep Election Data):
Users, please note that the colors have been flipped from traditional formats and the GOP is in blue (Dems in red).  What a change from 2008!  Walker annihilated vast regions that voted 60% of more for Obama in 2008, converting them to winners for his campaign.  The pattern is roughly the same, but shifted like 10 points right of the three-cycle averages.
And then he did it again.  In fact, his margin of victory was about 1.5 points bigger the second time around in his recall election.  I would call that noteworthy and a reason for optimism despite the leftward surge of an already left-leaning state in 2008.  The addition of Ryan to Romney’s ticket and the endorsement and campaign organization of Walker that is now being used by Romney’s campaign can’t hurt.

Swingers: Iowa

I covered a left-leaning swing state with a very homogeneous population last, now let’s look at a state that has come to define political warfare and launched many a political career – the thermostat of the nation and one of the swingiest of the swing states – Iowa.  Their primaries frequently select the most extreme of candidates (Obama over Biden and Clinton, Santorum over Romney, Huckabee over Romney and McCain, Kerry holding on barely over hot youngsters Edwards and Dean, who each flamed out for different reasons not related to a lack of Iowa enthusiasm, etc) and yet, county by county, this is not a land of extremes, as you’ll see.  No…it turns out that populism is dangerous, no matter what side of the social and political spectrum from which it emanates – tea-party demagogues like Michelle Bachmann (who faded due to bad campaign organization in Iowa, but was initially enormously popular) and evangelicals like Pat Buchanon and Bob Huckabee on the right, and extreme progressives and socialists like Obama and Edwards on the left equally pursue as their primary campaign strategy the ethos that they aren’t even really politicians – they’re one of us, hard working, ordinary and simple, with simple, hopeful ideas about serving the interests of the masses.  And Iowans love populists. They frequently don’t even realize that the people they’re choosing as their candidate are as extreme as they are – all they want is someone who went to all 99 counties and talked with them face to face.

But, as a group, Iowans are pragmatists in their daily lives and idealists in politics – go figure.  Let’s look at those 99 counties and see what picture they paint.

Major Population Centers:

  1. Sioux City (3% of state vote)
  2. Des Moines (14% of state vote)
  3. Cedar Rapids/Iowa City (~7.5% of the state vote)
  4. Dubuque (3% of the state vote)
  5. Davenport (5.5% of the state vote)

Slant grouping by county: average of the last three cycles
Slant grouping by county: 2008 election

The most representative county in the state is located between the states three eastern cities of Davenport, Dubuque and Cedar Rapids…its population is not large, but the more conservative suburbs of these left-leaning Iowa cities reflect a balance between union and government influence and rural and traditional values that pull at the state’s voters.  Cedar County is an ideal spot to watch on election night and you’ll have their votes before you have full votes from the cities…it should tell you something about republican turnout.

Obama’s statewide effort at early voting turnout paid big dividends in all parts of the state.  He didn’t just attack a specific region he thought might be vulnerable as he did in Colorado or New Mexico – his strategy was flawless.  Realizing that the entire state lacks severe biases and that the demographics cannot be used as a wedge in this region, his goal appeared to be to goose democrat turnout by a few points everywhere, shaving far-right counties into leaners, swing counties into left-tilters and urban areas into firm liberal hands.  This can be seen when we look at the slant group averages by cycle (as can the democrat turnout boost):

Turnout percentage by county slant grouping (from three-cycle averages) for 2000, 2004, and 2008 respectively:

1 6.63% 6.50% 6.33%
2 16.83% 16.56% 16.70%
3 22.96% 22.37% 21.79%
4 48.10% 48.83% 49.09%
5 5.48% 5.74% 6.09%

A boost in turnout in the D-Triangle (what locals call the region between Des Moines, Davenport and Dubuque) can be seen – note carefully that the lower percentages of the vote in swing and conservative counties were not caused by lesser turnout but by increased liberal turnout.

Slant ratios for the same groups for the last three cycles:

31.10% 29.66% 35.66%
44.05% 41.32% 47.71%
48.74% 48.51% 52.45%
54.22% 53.76% 59.06%
62.71% 63.71% 69.05%

We see a 4-6 point liberal swing in every grouping in 2008…those changes are not indicative of a general leftward trend in Iowa – Obama just got people to the polls better than most candidates and swayed independents.  The question is – will that tend continue?  Well we do have some information on that already. Early voting is well underway in Iowa and over 300,000 ballots had been requested as of October 12th (out of about 1.5 million votes normally cast in the state)…Democrats normally do well with early voting in Iowa, but Obama has taken that to a whole new level.  In 2008, at about this stage in the voting process, Obama had a party registration advantage of 56% D / 26% R / 18% I…this year, the requested ballots come out 47% D / 30% R / 23% I and independents are breaking for Romney by 5-8 points in all of the recent polls conducted in Iowa.  So…with the election about 20% completed, Romney is likely down by 13-15 points…but McCain was utterly BURIED by this time and he lost the state only by about 10 points.  So…there’s good reason to believe that Romney will be a bit more competitive here.  I am leaning toward projecting this state for Obama, but with momentum favoring Romney, that projection was more likely a couple of weeks ago.