USA 2012: WY, WA and Super Tuesday
After the primaries in Michigan and Arizona on February 28, the fascinating race for the Republican presidential nomination moved on to Super Tuesday’s seven primaries and three caucuses held on March 6. Side-shows caucuses of sorts were held in Wyoming (Feb 9-29) and Washington (March 3) between these two big sets of contests.
In my post on the last primaries, I compared this nominating season to a good TV show which returns to us almost every week with new intrigues, new twists and always a good load of suspense. In last week’s episode, Mitt Romney broke Rick Santorum’s momentum with a predictable landslide in Arizona and a close win in his home state but Santorum target state of Michigan. Mitt Romney surged to a pretty sizable lead in national polling over Santorum and second-tier rivals Newt Gingrich and Ron Paul. It seems very likely at this point that Romney will be the nominee, given the delegates he has amassed so far and his strength in late-voting WTA states which could place him over the top. However, his rivals are resilient and are unlikely to give him a free pass. The race could go on for quite some time still before Romney officially and formally seals the deal.
Between MI-AZ and Super Tuesday, Wyoming wrapped up its month-long county caucuses and Washington held a caucus on March 3. These caucuses, like – I think – all other caucuses held thus far, do not directly allocate delegates to the RNC in Tampa. These news-generating caucuses are only presidential preference straw polls with either no effect or a limited effect on delegate allocation, decided later in county conventions. The delegate projections created by media outlets based on the caucus results in these states thus vary wildly and are fairly inaccurate projections.
Wyoming (Feb 29) and Washington (March 3) caucuses
Wyoming (county caucuses)
Mitt Romney 38.99%
Rick Santorum 31.93%
Ron Paul 20.83%
Newt Gingrich 7.83%
Mitt Romney 37.65%
Ron Paul 24.81%
Rick Santorum 23.81%
Newt Gingrich 10.28%
Mitt Romney won a fairly comfortable victory in Wyoming’s month-long county caucuses (February 9 to 29) while in Washington he managed a comfortable victory over Ron Paul and Santorum. Washington was held in the wake of Romney’s post-MI momentum, which destroyed any chance for a Santorum victory. The delegate projections out of Wyoming indicate that Romney and Santorum both won roughly the same number, with Romney eeking out a narrow plurality. In Washington, Romney could have won between 30 and 34 of the state’s 43 delegates.
In Wyoming, the results indicated a fairly clear east-west split in the state’s GOP voting patterns. Mitt Romney dominated in the western part of the state, especially heavily Mormon Lincoln (75%), Uinta (65.7%) and Big Horn (70.4%) counties, all counties which showed turnout numbers heavier than the very low statewide average – only 2000 or so registered Wyoming Republicans turned out. Mitt Romney also carried the ski resort county of Teton (56.3%) fairly easily. In eastern Wyoming, he only carried Albany County (Laramie), and only with 35%. Rick Santorum and Ron Paul split the remaining counties in the east of the state. Santorum’s two major victories were in Laramie County (Cheyenne), in which he took 41.3%; and Natrona County (Casper) in which he took 38.6%. His other big win was in Goshen County, which seems sparsely populated but cast 146 votes, 65.8% of those for Santorum. Ron Paul won isolated plains county, including the old mining county of Sweetwater.
In Washington, Mitt Romney’s victory was helped in no small part by the heavily populated Seattle-Tacoma area. In King County, where most Republican voters tend to be affluent, educated Seattle commuters, he won 47% to Ron Paul’s 25%. In Snohomish County, another Seattle sprawl county, he won 42% to Paul’s 25%. In more exurban Skagit County (Mt. Vernon), Romney won 41% to Santorum’s 21%. In Pierce County, Tacoma and its suburban sprawl, he won 37.9% to Santorum’s 25.8%. Romney won eastern Washington’s main urban centre, Spokane, by a very narrow margin (30-29.8) over Santorum. He carried Clark County (Vancouver) with 37% against 28.5% for Paul. Vancouver is a fairly conservative urban area by PacNW standards, likely because it attracts the kind of residents who like low taxes (Washington has no income tax, neighboring Oregon has no sales tax). Romney took 43% to Santorum’s 25% in Benton County, home to the nuclear industry-driven Tri Cities.
Rick Santorum won Whatcom County (Bellingham, near the Canadian border) with 33% to Paul’s 28%. It may surprise, but it is likely that the GOP electorate in Whatcom County comes from Lynden rather than the liberal college town of Bellingham. And Lynden is an ultra-conservative Dutch Calvinist enclave, and those types of places have been Rick Santorum’s strongest locales thus far. Santorum also won three random eastern Washington counties where nobody lives. Ron Paul carried the four eastern Washington counties which border Canada, the coastal logging county of Pacific, two counties along the Columbia River and two counties in southeastern Washington. One of those counties, Whitman, is home to Washington State University (in Pullman).
Super Tuesday – Eastern Primaries (MA, VT, OH, VA, GA, TN)
Mitt Romney 72.09%
Rick Santorum 12.07%
Ron Paul 9.57%
Newt Gingrich 4.64%
Mitt Romney 39.79%
Ron Paul 25.49%
Rick Santorum 23.65%
Newt Gingrich 8.14%
Jon Huntsman 2.03%
Mitt Romney 37.95%
Rick Santorum 37.07%
Newt Gingrich 14.59%
Ron Paul 9.24%
Mitt Romney 59.52%
Ron Paul 40.47%
Newt Gingrich 47.20%
Mitt Romney 25.90%
Rick Santorum 19.56%
Ron Paul 6.55%
Rick Santorum 37.43%
Mitt Romney 28.09%
Newt Gingrich 24.18%
Ron Paul 9.11%
Super Tuesday – Western Primaries and Caucuses (OK, ND, ID, AK)
Rick Santorum 33.80%
Mitt Romney 28.04%
Newt Gingrich 27.48%
Ron Paul 9.63%
North Dakota (caucus)
Rick Santorum 39.74%
Ron Paul 28.07%
Mitt Romney 23.71%
Newt Gingrich 8.48%
Mitt Romney 61.59%
Rick Santorum 18.17%
Ron Paul 18.10%
Newt Gingrich 2.10%
Alaska (non-binding straw poll)
Mitt Romney 32.61%
Rick Santorum 29.03%
Ron Paul 23.96%
Newt Gingrich 14.15%
As the dust settled, it was clear that Mitt Romney eeked out a narrow win overall on Super Tuesday. The crucial state out of all 10 states which voted, the one which was most unpredictable and the one on which almost all candidates centered their attention on, was Ohio. And Mitt Romney, like in Michigan, was able to narrowly upset Santorum in the Rust Belt state, but only with 38% to Santorum’s 37.1%. A victory by the skin of his teeth, but still a momentum-maintaining win for Romney. Mitt Romney also emerged on top in Massachusetts, Vermont, Virginia and Idaho; where he was widely expected to win, and also won Alaska’s non-binding caucus straw poll. Rick Santorum won Tennessee, Oklahoma and North Dakota. Newt Gingrich won his home state of Georgia.
There are two ways to look at Romney’s victory in Ohio. On the one hand, Romney supporters will likely perceive it as a narrow victory but a victory nonetheless for Romney – which makes his likely nomination even more certain – in a state demographically favourable to Santorum (which isn’t quite true). On the other hand, a lot of people will probably see Romney’s victory as an underwhelming victory in a state where he outpsent Santorum 5 to 1 but was only able to beat him by less than 1% despite blowing him out of the water on the money race side of things.
Mitt Romney’s victory in Ohio does not seal the deal for him just yet, but it indicates that Romney’s route to the eventual crowning moment will be a little bit shorter than it would have been if he had lost to Santorum in Ohio. Mitt Romney’s delegate advantage increased on Super Tuesday, and he now has roughly 400 delegates, with 1,144 needed to win the nomination. Realistically, this lead is likely insurmountable for either Santorum or Gingrich unless one of them dropped out in favour of the other and was able to gather tons of momentum and cash to challenge Romney in his new firewall: WTA states in the Northeast plus California. However, Santorum and Gingrich are both proving to be resilient fight and it is tough to imagine either of them dropping out this week. Gingrich has little to lose in this contest, and Santorum seems to be in no mood to just give up and give Romney the nomination. Romney could emerge as the official nominee by the end of April or mid-May. Only Gingrich dropping out and giving Santorum the chance to build a conservative coalition could rejig the race, but even then it could be too late. And Gingrich won’t drop out for a week at least.
Romney’s victory on Super Tuesday is murkier than headlines indicate. He has failed to overcome his “Southern problem” or “conservative problem”. He lost to Santorum by fairly consequent margins in Tennessee and Oklahoma, despite a fairly divided conservative electorate in these conservative Republican bastions. In Ohio, as our geographic analysis will show, Mitt Romney – like in Michigan – won because of the votes of GOP voters in big-city Democratic strongholds and swing-vote suburbs, but lost to Santorum in the traditional Ohio Republican strongholds. All this indicates that while Romney will win the nomination, he will do so with a conservative base which is fairly unexcited about him to say the least and generally lukewarm towards his candidacy. John McCain faced a similar problem in 2008 but his selection of Sarah Palin as his Veep turned matters around for him as the same conservatives who had shown reluctance towards McCain were energized by the Palin pick. Romney could resolve the issue in a similar fashion, but at this point in time, he faces an uphill battle to gain the confidence of these voters. The overall results also indicates that Romney could struggle in the general election against Obama in working-class areas, but at the same time do well in suburban areas.
State-by-State Analysis: Exit Polls and Geographic Analysis
Massachusetts was the most boring contest of the night: Romney won 72% of the vote and won all 38 delegates which were up for grabs. With such a margin, you could think that Massachusetts is full of Mormons! It doesn’t actually have lots of Mormons besides Romney, but it does have other things: it is Mitt Romney’s adoptive home state – where he served as Governor between 2003 and 2007 – and its Republican electorate tends to be moderate, affluent, educated suburbanites. A huge landslide is what happens when a favourite son candidate named Romney is the only ‘moderate Republican’ on the ballot. The fact that the other candidates totally ignored the state also explains stuff to some extent.
Romney’s win in Massachusetts in 2008 was nothing to write home about – he beat McCain by only 10 points in his home state – but that was largely because McCain, favourite son effect erased, was a much better candidate for Massachusetts GOPers than the conservative Romney of 2008.
Exit polls, of course, are boring. Romney won 80% with those aged 65 or over, a group which made up 29% of voters. His support was still kind of graduated by income, but not as perfectly as before. He won 73% with the top 10% – those making over $200k, but took 77% of those 31% with an income between $100 and $200k.
Independents were 51% of the electorate and moderates/liberals were 49% of primary voters. Romney did better with registered Republicans (78%) than with independents (69%, Paul took 14%), and won 72% support from moderates against 64% support from ‘very conservative’ voters (15%). However, he won the most support – 76% – from somewhat conservative voters. Romney won 69% support among the 51% of voters who said that so-called RomneyCare – the state’s healthcare law passed by Governor Romney and later the blueprint from ObamaCare – went too far. This might explain why attacks on RomneyCare don’t seem to stick to Mitt: voters tend to disassociate the two or at least don’t consider Romney responsible for it. Romney won 82% support from the 43% who said that his ties to Massachusetts mattered a lot or a bit to them.
On a geographic basis, Mitt Romney received the most support in and around Boston in eastern MA. These counties (Essex, Middlesex, Norfolk, Plymouth) are made, largely, of moderate, highly educated and very affluent suburban communities. Plymouth is slightly more exurban and less affluent, but Republicans here – and there are quite a few by local standards – care a lot about taxation and such stuff. Romney won 72% in Brookline, 80% in Needham, 75% in Newton, 83% in Wellesley, 74.6% in Framingham, 72% in Waltham, 77.8% in Milton, 72.7% in Quincy, 75% in Weymouth and 82% in Duxbury. In this densely populated region, his only more underwhelming performances were in less affluent, more working-class areas such as Somerville (59%, 23% for Paul), Chelsea (54%, 27% for Paul) or liberal college towns like Cambridge (59%, 21% for Paul). He won 69.6% in Boston.
Romney also performed well around Cape Cod (Paul did well in Provincetown), with 75% in Barnstable, for example. He didn’t do as well in the declining post-industria cities of Fall River (55%, 21% for Santorum), New Bedford (54%, 22% for Santorum), Springfield (58%, 23% for Santorum), Holyoke (59%, 20% for Santorum) or Chicopee (59.5%, 23% for Santorum). His worst performances, however, were in the Berkshires, where he sometimes hovered below 50% and rarely did better than 60%. This is a Vermont-like liberal stronghold, but Santorum and Paul did fairly well. Interestingly, Paul didn’t do spectacularly well in college towns such as Amherst (21%) or Adams (13.7%).
Vermont has shifted away from the Republican Party in droves in recent years, but we usually assume that those who remain in the Vermont GOP tend to be moderates. Based on this assumption, Romney should have done very well in Vermont. But his victory, with 39.8% to Paul’s 25.5% is not the landslide we might have assumed. When we think about stuff in more detail, it makes sense. Moderate or liberal Republicans are endangered species, but the idea that Romney dominates them has not really been proven. Mitt Romney won New Hampshire by the margin he did largely because of more conservative or libertarian affluent Boston suburbanites, while he did poorly in western NH, which most resembles Vermont in political and demographic terms. Vermont is not particularly well-off, and it certainly doesn’t have those New Hampshire-type very affluent suburban voters who are Romney’s strongest backers. It also has a fairly anti-establishment, independent streak which is hard to quantify or even observe in specific elections, but which can rear its head from time to time.
The exit polls prove that this observation is true: only 3% of Vermont voters made over $200k and only 17% made over $100k. Romney won that top 17% with 42%, while Ron Paul won the poorest 13% of voters (income under $30k) with 37% to Mitt’s 32%. Independents were 40% of the VT primary electorate, but for the first time this year, conservatives were outnumbered by moderates-liberals in Vermont: only 47% of the voters were conservative. Paul won independents, 38-31 over Romney, while Romney won Republicans 51-25 over Santorum. Santorum won the 19% who identified as very conservative, while Paul lost the moderates and liberals by only one point to Mitt (34 vs. 35).
Mitt Romney did best around Burlington. He won Burlington proper with 37% to Paul’s 29%, but did far better in the kinda-suburban towns which surround Burlington and which tend to be slightly wealthier. He won 51% in affluent Shelburne, 47% in South Burlington, 43% in Colchester, 42.6% in Essex, 42% in Jericho and 45.8% in Williston. Outside Chittenden County, Romney also did well in Rutland (45%), Bennington (38.5%), Brattleboro (39.8%) and especially the affluent ski resort of Stowe (48.6%). Ron Paul did really well in the Northeast Kingdom (Essex, Orleans and Caledonia counties) but also most of Lamoille County and inland Franklin County. Rick Santorum took a few towns here and there too, including Highgate on the Canadian border. The towns won by Paul or Santorum are largely sparsely populated rural small towns in the Green Mountains, where voters are pretty poor and portray Vermont’s independent, anti-system streak fairly well. Ron Paul also won Marlboro, a college town in southern Vermont, but there certainly isn’t any college town rule in the results. Santorum won Putney; Romney performed strongly in Northfield, Middlebury, Norwich and Hartford.
Ohio was the race which everybody was interested about. It was the most competitive contest of all 10 states which voted on Super Tuesday, and it was where Romney, Santorum and Gingrich focused their strengths. Rick Santorum polled very well in Ohio right up to his loss to Romney in neighboring Michigan, which allowed Romney to close the race down to a statistical tie. Romney outspent Santorum, whose campaign was so disorganized it failed to qualify for a full slate of delegates in each CD, by a 5-to-1 margin. Yet, Romney, unlike in Florida, was unable to use this money advantage to blow Santorum away. It has often been said that Ohio is demographically favourable to Santorum, even moreso than Michigan. This is not quite true – its demographic makeup is either as favourable or slightly less favourable to him than Michigan was. It has more Catholics, less Evangelicals and no Dutch Calvinists. In one of the closest races of the primary, Romney won Ohio with 38% to Santorum’s 37.1%. Newt Gingrich, who had focused on the state to an extent, won only 14.6%, a bit less than what the polls had given him. I discussed the two ways of interpreting this and the significance of this in more details above. I will now look at how Romney won, and why I subscribe to the view of Romney’s Ohio win as underwhelming.
The exit poll provides interesting information. Firstly, in terms of age groups, Santorum won all age groups besides those 65+, which Romney won by a crushing 47-31 margin over Santorum. In addition to what we have observed since day one about Romney’s support increasing as one’s personal income increased, we can add to that another strong correlation: Romney’s support really increases as one gets older. The income correlation was still there, of course, but interestingly the correlation was not quite perfect. Romney, of course, won those making over $200k with 53% to Rick’s 24%, but his worst income group were lower middle-classes ($50-100k), where he got just 32%, and not the lowest 15% (those making under $30k), where he tied Santorum at 35-35. Santorum stood at 43%, his strongest result, with those making $50-100k.
In ideological terms, 66% of primary voters were conservatives, and Santorum won that large group with 41% to Mitt’s 35%. With the third of voters who were very conservative, he won 48%. Moderates or liberals, a third of voters, backed Romney 43-29 over Santorum while also giving Paul his best result (13%). Romney won Republicans (41-37) but Santorum won both independents (26% of voters, 37-31) and Democrats (5% of voters, 47-27).
47% of voters were evangelical or born-again Christians, and they picked Rick by a 17-point margin (47-30) over Santorum. A third of voters were Catholic, and Santorum lost his coreligionists 44-31 to Romney while winning Protestants by a narrower 41-39 margin over Romney. Santorum had already lost fellow Catholics in Michigan and Iowa. It is interesting that there is absolutely no ‘Catholic vote’ for a fellow coreligionist. From a psephological aspect, I think this goes a long way to explain the general nature of Catholic voting patterns in the United States. If one seeks an explanation for this rather interesting element of the exit polls, it might be because social conservative and Evangelical/Christian right voters tend to be disproportionately Protestant rather than Catholic, or that Catholic voters tend to care more about economic issues than culture war/wedge issues such as abortion or gay marriage. Many American Catholics have moved away from their Church’s traditional conservative position on those issues and tend to be quite secularized despite claiming a Catholic faith or tradition.
Mitt Romney won his narrow victories on the back of big margins in late-counting big cities and inner suburbs. Romney won all of the main urban counties including Cuyahoga (Cleveland), 48.7-29.6; Summit (Akron), 43-34.3; Franklin (Columbus), 40.7-36.1; Hamilton (Cincinnati), 48.9-29; and Montgomery (Dayton), 39.7-31.4. The only major city he lost is working-class Toledo (Lucas County), in which he took 36.6% to Santorum’s 37.8%. Republicans in these traditionally Democratic counties tend to be affluent, educated and more suburban than the county’s population as a whole. Cuyahoga County certainly includes some very affluent suburban places, besides Democratic inner-city Cleveland. Columbus and Cincinnati are also largely white-collar cities with big corporations and affluent GOP-leaning residents. Cincinnati (Hamilton County) is a conservative metropolitan area by almost all standards, perhaps because of its large German Catholic population or particularly rock-ribbed GOP suburbs filled with affluent voters.
Besides the big cities, Romney also won their highly-educated and affluent suburbs or exurbs. He won 41.6% to Santorum’s 34.6% in Warren County in suburban Cincinnati and 41.9% to 34.4% for Santorum in next-door Butler County, an affluent exurban-suburban area. In suburban Columbus’ Delaware County, he won 42.3% to Santorum’s 35.9%. In the greater Cleveland area, he crushed in very wealthy Geauga County with 45.7%, but also carried slightly less affluent suburban Lake County (43.5-32.2) and exurban Portage County (39-35) and Medina County (40.8-34.7). He also won Erie and Lorain Counties, whose GOP voters tend to be suburban or exurban and fairly wealthy.
In these close races, people like to cling to random things and sensationalise about how candidate x owes his victory exclusively to those things. In this race, you can say that Romney won because he won the urban counties big, because he won Catholics or because he won working-class Catholics. I don’t like sensationalising in such ways, but from one point of view, Romney ironically won, in part, on the back of his narrow victories in working-class Catholic areas. In Youngstown-Warren, a low-income and working-class post-industrial urban conglameration, Romney beat Santorum 37-34.5 in Mahoning County (Youngstown) and 35.8-35 in Turnbull County (Warren). These post-industrial counties have a big Catholic population of Eastern European, Irish or Italian descent in large part. We should perhaps re-evaluate all the stuff which has been written about Santorum’s particular appeal to working-class voters in the Rust Belt. His appeal in older, urbanized manufacturing and post-industrial cities, which tend to have a large Catholic electorate, has been fairly limited. He did win Toledo and Flint, but fairly narrowly; but he lost Saginaw, Bay City, Macomb County and now Youngstown-Warren. His Rust Belt populist appeal seems to be working out in more rural, less big-city, less solidly Democratic working-class areas.
Rick Santorum won the rest of the state. The rest of the state includes very conservative rural ‘Corn Belt’ counties in western Ohio, which has a large rural German Catholic population which Santorum likely won; Protestant Evangelical and low-income voters in the corridor between Akron and Columbus; working-class Rust Belt areas in the Ohio River valley; and culturally Southern voters in southeastern Ohio (which includes a bulk of counties with a plurality of ‘American’ ancestry residents). In the Appalachian white working-class (mining, manufacturing, steel) counties of the Ohio River valley, an area where Obama had really struggled in 2008, Romney is roughly in the same boat as Obama was. He failed to break 30% in a handful of counties in this area, including Jefferson County where Santorum won 57.7%. I’m not sure what’s up in Athens County (60% for Santorum, 19.6% for Romney) – it could be an error – but it seems like it may be another case of Alachua County, Florida – a liberal county with a big college town which leans heavily to the left, but with a Republican electorate which is extremely conservative.
Virginia’s primary was a rather bizarre affair: only two candidates – Mitt Romney and Ron Paul – gathered the required signatures to appear on the ballot, leaving Santorum and Gingrich off the ballot in the state where both of them are currently registered to vote. The result was a primary basically conceded to Romney, but also a chance to measure how Paul – the least popular of the anti-Romneys amongst the social conservative/right-wing GOP crowd – could measure up to Romney in a contest where he was the only anti-Romney. In the end, Romney won, of course, taking 59.5%, but Ron Paul’s 40.5% was a very strong showing for him. Virginia certainly isn’t prime Paul territory and I think he would have had trouble breaking 10% in a normal primary, so he obviously took quite a number of votes from the anti-Romney crowd, which is likely pretty strong in Virginia which is at least half-Southern in its makeup. Virginia is not entirely relevant, as turnout was low and the Paul base was likely very motivated, and the anti-Romney crowd didn’t turn out en masse, but I still think it speaks volumes about Romney’s base problem that he only won 59.5% of the vote against a guy who is widely considered to be unelectable and who is the only contender who hasn’t won one state thus far.
Exit polls reveal how the primary electorate was small and hardly representative of a normal VA GOP electorate. 34% were moderates or liberals, which seems high for Virginia, and only 44% of voters were Evangelical or born-again, which seems low for Virginia. Otherwise, Romney won older voters (83% with those 65+), Paul won won those 17-29 (61%) and 30-44 (63%). Ron Paul did much better (48%) with those earning $30-50, the lowest income group to be quantified, but lost heavily (64-36) to Romney with those voters making over $100k.
Paul won independents, a third of the electorate, with 64%, but lost Republicans 73-27 to Romney. He tied Romney with the 34% who described themselves as moderates or liberals, and won 36% support from the very conservative voters (32%).
Ron Paul actually won a few counties, quite a few of them too. He won a fairly bizarre string of them in southwestern Virginia, all of which were won by Huckabee over McCain in 2008. One of these counties, Montgomery County includes the liberal college town of Blacksburg (Virginia Tech), but I’m tempted to attribute these victories to a conservative anti-Romney vote, although one which seems fairly limited because Paul certainly didn’t have Huckabee’s appeal in southwestern Virginia, the most Dixie-like region of the state.
Paul also won Lynchburg (51%), a conservative college town which includes the Christian right’s Liberty University; the liberal college town of Charlottesville (52%), Manassas Park (53%) and random Buckingham and Warren counties. He also proved popular in black-plurality Norfolk (50.6%), Portsmouth (51.5%), Surry County (53.5%) and Charles City (52.2%).
On the other hand, Romney blew Paul out of the water in Richmond’s affluent suburbs: 63.9% in Henrico County, 67.3% in Goochland County, 62% in Chesterfield County, 57% in Powhatan County and 57.3% in Hanover County. In Richmond proper, Paul took 48.5%. Romney also dominated in NoVa, where Republicans tend to be of the very affluent and highly educated demographic so favourable to Romney. He won 62% in Loudoun County, 65.3% in Fairfax County, 60.8% in Prince William County, 67.6% in Alexandria and 64.6% in Arlington. Romney also did very well – breaking 70% in two counties – in the Chesapeake Bay region, specifically the Northern Necks, where I assume you find a fair number of affluent retirees in the small coastal resort communities.
Georgia is Newt Gingrich’s kinda-home state, and certainly the state where his base is strongest and where he has maintained strong support despite his campaign’s descent into the near-abyss since Romney handily defeated him in Florida over a month ago. Santorum seemed to be in a position to give Gingrich a bit of a race, but Gingrich had a mini-surge of sorts in Georgia following Santorum’s momentum-crushing loss in Michigan a week ago, and the conservative vote united around Gingrich and abandoned Santorum. The result was a strong victory for Gingrich in a delegate-rich state, taking 47% to Romney’s 25.9% and denying Santorum, who won only 19.6%, a chance to get delegates out of the state.
Romney had won 30.2% of the vote in Georgia in 2008, meaning that he actually did better in 2008 than in 2012 in Georgia. Newt Gingrich’s landslide victory carries us back to the days of South Carolina back in January, but we shouldn’t delude ourselves into thinking that Georgia will resuscitate his fledgling candidacy. It was a favourite son victory, about as relevant as Howard Dean winning Vermont in 2004 or John Edwards winning North Carolina that same year. It wasn’t really a Southern candidate victory, as his performance in Tennessee shows.
Newt Gingrich dominated nearly every single demographic. He polled strongly even in Romney’s core 65+ constituency, won all income levels besides those who make over $200k (Romney won them by one, 39-38). He trounced Romney 53% to 19% among very conservative voters (39% of voters), a group where Santorum actually ran a distant second with 25%. He even won moderates, with 40% to Mitt’s 28%. 64% of voters were Evangelical, and he took them with 52% against 24% for Santorum. On the other hand, he only defeated Romney by one (38-37) among the minority of voters who were not Evangelicals. He also lost Catholics, 12% of voters, by four points to Romney (38-34), despite being Catholic himself. Even among those 45% of voters citing one’s ability to defeat Obama as the top candidate quality, he beat Romney by 10 (48-38). 61% of voters say that Gingrich’s tie to Georgia didn’t matter, but all these numbers indicate that Gingrich got a big favourite son vote in Georgia. I mean, 39% of voters considered him the candidate most likely to defeat Obama in November…
Romney had done well in the Atlanta metro area in 2008, and it was where he did best this year again. He won convincingly in Fulton County (Atlanta), 45.6% to Gingrich’s 33.2%. He also carried neighboring black majority DeKalb County (39.2-35.7). Fulton County should actually be two counties; a southern part (Atlanta) which is heavily black, low-income and very Democratic and a northern part (Sandy Springs, Alpharetta, Roswell, Milton) which is heavily white and includes some of the wealthiest areas in the Deep South. Romney won big there in 2008, and won big again in 2012. In Fulton County, the GOP electorate is not as conservative as the one found outside the Atlanta metro. The Republican parts of DeKalb County, few and far between, are also very affluent. However, Gingrich beat Romney in four suburban/exurban Atlanta counties Romney had carried in 2008. In Cobb County (Marietta), more exurban despite some very wealthy inner suburban areas, Romney lost 43.7% to 32%. He won only 28.8% in Gwinett County, 20% in Clayton County, 28% in Forsyth County, 28.3% in Fayette County; all exurban counties. Gingrich used to represent GA-06, which in his time covered northern Fulton County and parts of Cobb and Cherokee counties. He likely won his old district’s old boundaries convincingly.
Romney’s only other victory was in Chatham County (Savannah), with 39% to Gingrich’s 35%. Besides Savannah, it includes some very affluent coastal island resort communities (Skidaway Island, Wilmington Island).
Newt Gingrich swept the rest of the state, similarly to South Carolina. He won the Black Belt areas, generally supportive of establishment candidates in presidential primaries (McCain won it in 2008); the white rural areas of southern Georgia; Atlanta’s exurbs and northern Georgia. Like in South Carolina, Gingrich was able to build a favourite son coalition made up of more populist and Evangelical Upstate (northern Georgia) voters and the more patrician-tradition and pro-establishment conservatives of the coastal plains and Midlands. Gingrich took most of Georgia’s mid-sized urban and suburban areas. He won Bibb County (Macon) with 46.9% to Mitt’s 26.8%, Muscogee County (Columbus) with 40% to Romney’s 29.7% and Richmond County (Augusta) with 40% to 28.9%. Gingrich won Clarke County (Athens, a college town) by a smaller margin: 39% to 30%, with Paul pulling in 13.5% in fourth place. Romney performed slightly better in these urban areas the more affluent Colonial Coast (Golden Isles region), but failed to carry Glynn County (Brunswick) which he had won in 2008. However, in the bulk of rural Georgia – north and south – he was badly trounced, rarely breaking 20% and often placing third behind Rick Santorum – whose support is closely correlated with the map of white Evangelicals.
As we found in Jacksonville, Florida; Romney’s affluent suburban base is rather limited in the Deep South. He had done well in those conservative Dixie suburbs in 2008, but that was when he was the non-Evangelical conservative contender rather than the blander establishment moderate in the race. Southern suburbs are more conservative (tending to be the most Republican counties in the state) and less ethnically diverse than most of their northern counterparts, which are more receptive to more moderate candidates such as Mitt Romney. White flight is also a major phenomenon in a lot of the newer Southern suburban counties, and the type of voter that such suburbs contain are hardly favourable to him. Romney lost Savannah white flight Effingham County 41-23 to Gingrich, placing third behind Santorum. It is hard to quantify, but Romney has shown that he has only very limited appeal to Southern voters in newer suburban or exurban areas, his Southern suburban strength being really just concentrated in the wealthiest of the older inner suburbs.
Tennessee emerged as the second most competitive Super Tuesday contest after Ohio. A Southern state where Gingrich lacked a favourite son appeal, it was to be the first test for Rick Santorum’s ability to win in the Deep South despite not being a Southerner in a race which features a Southerner (Gingrich). Until the final few days, it seemed as if Santorum would win Tennessee easily, but after Michigan, his numbers fell and Gingrich’s numbers rose some. The division of the conservative vote between Santorum and Gingrich gave Romney the chance to creep up the middle and win what could be a symbolic victory in the South. It did not come to be. Santorum won 37.4% to Romney’s 28.1%, a decisive victory. Newt Gingrich performed fairly strongly with 24.2%, but this was only good enough for an unremarkable third place showing – in a Southern state bordering Georgia no less. Since Nevada, Gingrich has failed to come second or better in any state except Georgia. That shows how moribond his campaign is at this point.
Tennessee’s GOP electorate is conservative – it voted for Huckabee over McCain in 2008 – but at the state level it has tended to support moderately conservative establishment candidates like Bill Haslam, Bob Corker or Lamar Alexander over insurgent conservative candidates. Romney faced an uphill fight in Tennessee, but it would not have been impossible for him to win if he had proved to have a larger base appeal.
Santorum swept most demographic categories in the Tennessee exit poll, leaving Romney to his core demographic stregths: older voters (65+, he won them 34-31) and the wealthiest (those making over $200k, he won them 47-26). Santorum did better with middle-aged voters, as well as poorer and lower middle-class voters.
The electorate was overwhelmingly conservative, at 73% identifying as conservatives including 41% who were ‘very conservative’. Republicans made up 68% of voters, independents made up an additional 27% and 5% of voters were Democrats. Santorum won Democrats (41-21) and independents (38-25) by larger margins than he won Republicans (38-29 over Romney, Gingrich pulling 27%). With the very conservative voters, Santorum won 48% to Newt’s 27% and Romney’s paltry 18%. Romney, however, won ‘somewhat conservative’ voters by two (35-33) and moderates by five (33-28). 73% of voters were Evangelical, a group which Santorum won with 42% to Gingrich’s 25% and Mitt’s 24%. Romney still dominated with those who felt one’s ability to beat Obama was the most important quality (40-32 over Gingrich, Santorum in a poor third with 25%), and 43% of voters saw him as the candidate most likely to win in November. But, on the other hand, a full 49% of voters felt that Romney’s positions were not conservative enough.
As we found in Georgia, Mitt Romney’s base was rather limited. He had won a handful of counties in 2008, when he had won 23.6% in Tennessee, but this year he won only three counties. Two of them were in the Nashville area. He took 33.1% to Santorum’s 30.9% in Davidson County (Nashville) and 35.8% to Santorum’s 32.5% in Williamson County (Franklin, south of Nashville). Republicans in Williamson and Davidson counties, which include suburbs of the like of Forest Hills, Oak Hill and Brentwood tend to be the most affluent voters in the state – Williamson is the wealthiest county in the state. Romney also won, more randomly, Loudon County (36.2-34.6) which seems to include some more affluent suburbs of Knoxville in eastern Tennessee. However, Romney, like in Georgia and Florida, was unsuccesful in the newer, solidly Republican upper middle-class exurbs or outer suburbs of Nashville and Memphis. He had won exurban Nasvhille’s Rutherford, Sumner and Wilson counties in 2008; this year he lost them all. He lost 41-24 in Rutherford, 38-27 in Sumner and 40-24 in Sumner (placing third behind Gingrich). Romney also lost Shelby County (Memphis) 37.4-34.2 to Santorum. White flight is more pronounced in Memphis’ otherwise affluent suburbs included within Shelby County.
Rick Santorum swept the rest of the state save for one (or two? there are differences between sources) in eastern Tennessee which voted for Gingrich. Santorum was able to put together a coalition composed of East Tennessee Hill Country, Middle Tennessee and West Tennessee. He narrowly won Knox County (Knoxville) with 34.5% to Mitt’s 33.7% and prevailed in Hamilton County (Chattanooga) with 31.5% against 28.9% for Romney. Despite being hilly and historically very much opposed to the patrician plantation owners of West and Middle Tennessee, East Tennessee’s ancestrally Republican (Unionist since the Civil War) electorate has, unlike the Upland voters in South Carolina, usually favoured establishment candidates. Gerald Ford won most of the region in 1976 against Ronald Reagan and John McCain did fairly well against Huckabee there in 2008. This year, Mitt Romney did fairly well too in East Tennessee, but Santorum did well enough to sweep it in its quasi-entirety (despite strong Gingrich showings). He won, however, his best results (over 40%), in the more purely Dixie plain country of Middle and West Tennessee, where candidates of the populist/Huckabee variety do well.
Oklahoma is a very conservative state, but tends to have fairly erratic presidential primary voting patterns for both parties. John McCain narrowly defeated Mike Huckabee in 2008, in a map which revealed a split between the more Southern parts of the state and the Midwestern parts of Oklahoma. Rick Santorum, unlike Huckabee, has proven to be more than just a sectional candidate and has real appeal to both Southern and Midwestern conservatives; while Romney doesn’t have McCain’s appeal to Midwestern conservative voters. Oklahoma was always going to be a slam-dunk for Santorum after his post-CO/MN surge. He won a fairly poor 33.8% to Romney’s 28%, hurt in good part by Newt’s very strong showing: 27.5%. I find it amusing that Romney basically won the same percentage in both Tennessee and Oklahoma. Can we assume that Romney’s Southern base of sorts is 28% of the vote in a three-way contest?
The exit polls are somewhat interesting. Gingrich won men, but lost women to Santorum (and Romney) by a big margin. He actually performed very strongly with the 65+ crowd (40%), while Romney did meh with those voters, usually his top demographics (only 29%, he did better with those 30-44). Romney did, however, win the wealthiest voters: he took those earning over $100k by a 10 percent margin (40-30) over Santorum and tied Santorum among those earning $50-100k. Santorum (39%) and Gingrich (35%) both performed best with those earning less than $30k. Evangelicals were a full 72% of the electorate, and Santorum won them by 10 (37-27) over Romney and Gingrich.
Predictably, this being Oklahoma, conservatives made up 75% of the GOP primary electorate, including 47% who were “very conservative”. Santorum won both groups, the latter by a crushing 40-32-21 margin over Gingrich and Romney and the former by 10 over Gingrich (38-28, 25% for Romney). Romney won the quarter of voters who were moderates, 38-28 over Gingrich with Santorum pulling just 19%. Santorum still had major problems convincing the 40ish percent of voters who feel that a candidate’s ability to beat Obama is top candidate quality; he won just 18% with them.
Mitt Romney, again, saw his appeal concentrated heavily in urban areas. He won Oklahoma County (OK City) 34.5% over 30.6% for Santorum, but he lost Tulsa County (Tulsa) 32.3% to 28.8% to Santorum – and placed a close third behind Gingrich (29.6%). He lost Comanche County (Lawton) 30.5% to 35.6% for Santorum. Romney’s other win was in Payne County, home to the college town of Stillwater, which he won 31.3% to 28.2% for Santorum. Romney failed to prevail in OK City’s two main suburban counties; Canadian County (lost 34.8% to 27.7%) and Cleveland County (Norman, lost 33.1% to 30.3%). He placed a poor second or third in the more exurban counties of OK City and Tulsa.
Newt Gingrich won a few counties, in a way which is so random that it is hard to explain. He won around Enid and Woodward in Midwestern northwest Oklahoma, did well around Tulsa but fairly poorly in Little Dixie. Santorum won the rest of the state, with appeal to both Midwestern and Southern-like areas of the state. He did well in Little Dixie, but also did very well in the very conservative Oklahoma Panhandle, which is very Midwestern.
Idaho is a conservative state, it is a caucus state; so based on those two factors, Romney shouldn’t have done overwhelmingly well. Indeed, some observers were fairly conservative about his chances in Idaho. But Idaho is the second most heavily Mormon state after Utah, with some 26% of its population being Mormons, heavily concentrated in eastern Idaho – or “northern Utah”. Given how solidly Republican the Mormons are, and how motivated of a base they are for Romney this year, we can estimate that Mormons made up at least half of the Idaho caucus electorate this year, if not close to 55-60% of the whole caucusgoers in Idaho this year. Thus, predictably, Mitt Romney won Idaho easily, taking 61.6% to Santorum’s 18.2% and Paul’s 18.1%. The map is all shaded in with over 50% shades because ID caucuses are run with an intricate recaucusing system, voting in each county continues through successive ballots until a candidate receives a majority or only two candidates remain (at which point a final ballot is taken). Any candidates placing below 15%, plus the bottom remaining candidate are eliminated each round. This explains why in some counties, when looking over results in details, you will find some straight two-way contests excluding two of the other candidates – like Romney – because the others failed to qualify for the final ballot.
There were no entrance polls for the caucus in Idaho, unfortunately, but it would have revealed some interesting things about Mormons vs. non-Mormon Protestants in Idaho’s GOP caucus electorate. We can safely say that Romney like won some 90-95% of the vote with Mormons, but at the same time lost the non-Mormon minority by a sizable margin to either Santorum or Paul. Our map of the result confirms this, by highlighting a major fault line between eastern and western Idaho/the Idaho Panhandle. In eastern Idaho, which is very heavily Mormon (like Utah), Romney killed. 79.5% in Bonneville County (Idaho Falls), 79.2% in Bannock County (Pocatello), 78% in Teton County (Mormons-n’-ski bunnies). In the smaller, rural counties of the region, he broke 80% with ease. In tiny and heavily Mormon Franklin County (which we can take as a good example) he took 86.1%. It is interesting to point out that Paul often did comparatively well in Mormon country, breaking 10% in a few counties including Franklin County. Some stuff has been written about Paul’s appeal with Mormon voters, based on his constitutionalist principles which seem to appeal to some Mormons not enamoured by their coreligionist Mitt Romney.
Mitt Romney carried Blaine County (Ketchum-Sun Valley) with 60.5%, likely because the ones who aren’t Democrats there are probably Mormons or at least affluent ski resort Republicans. Similar comments can be made about the Boise-Nampa area, which he won on the back of a Mormon base mixed in with suburban affluence. He took 51.8% to Santorum’s 22.8% in Ada County (Boise) and 51.8% to Santorum’s 32.7% in Canyon County (Nampa).
On the other hand, Mitt Romney failed to carry a single county in the Panhandle, heavily non-Mormon, though he did do well in Nez Perce County (Lewiston) and Shoshone County, where low GOP turnouts leads to a strong Mormon base in the GOP caucus-going electorate. There remains a fairly strong anti-Mormon sentiment in these parts of western Idaho, which despite being roughly as conservative as Mormon country, have little else in common politically. Romney often failed to qualify for runoff viability in a handful of counties in the Panhandle. Ron Paul won easily in Latah County (Moscow, a college town) with 52.3% to Romney’s 20.2% and narrowly beat Romney in the runoff in more working-class Nez Perce County (Lewiston) with 50.6%. Santorum, however, did win most of the Panhandle’s working-class belt, taking 63.8% in Lewis County, 64% in Clearwater County, 54% in Shoshone County and 50.9% in Benewah County. He won the region’s main urban centre, Coeur d’Alene in Kootenai County with 57.6% in a runoff against Paul.
North Dakota, a caucus state, went for Romney on Super Tuesday in 2008, but it was a tough state to predict. Some were reluctant to give the state to anybody else given that Romney won it, while others claimed that Santorum’s success in surrounding Plains state guaranteed him a win in conservative North Dakota. They ended up being right, as Santorum easily won with 39.7% to 28.1% for Ron Paul. Mitt Romney placed third with 23.7% in a state which went to him with 35.7% in 2008. We can now ascribe Romney’s win in 2008 to the “conservative caucus” effect, a conservative crowd of caucus-goers which turns out for the ‘pure’ conservative candidate in the race. Romney’s advantage in caucuses was overwhelming in 2008, and while he hasn’t lost it entirely this year, his caucus performances are underwhelming more than anything.
There were no entrance polls in ND, and the results were only reported by state house district, which Google Politics was good enough to give us. Results by house district are both less detailed in rural areas where districts cover many counties, and more detailed in urban areas where house districts cover only parts of a single larger county. Rick Santorum swept the bulk of rural North Dakota, his lowest showing in rural North Dakota coming from HD-9, a predominantly Native American district where he polled third with all of 13 votes against 15 votes apiece for Paul and Romney. In rural ND, Ron Paul performed best in the more hilly areas to the west and north of the Missouri River, including the Badlands and Little Missouri Grasslands. Santorum did better in the traditional Plains region of rural ND.
Santorum also prevailed in the state capital, Bismarck, losing only an affluent northern suburb to Romney, though Paul did well in the city’s small core. Romney won Minot with 44% to Santorum’s 25%; the presence of Minot AFB likely explains Romney’s advantage. Santorum seems to have narrowly prevailed in Grand Forks, although both other candidates won a district. Ron Paul won the college town of Dickinson with 36.7% to 35.8% for Santorum. Ron Paul dominated in Fargo, the state’s largest city and home of NDSU. Santorum only won two districts, which seem affluent, south of downtown Fargo.
Alaska can take the prize for most erratic voting patterns in GOP primaries. Steve Forbes almost won the state against George W. Bush in 2000, Pat Buchanan won it in 1996 and Pat Robertson won in Alaska in 1988. In 2008, Mitt Romney carried the Alaska caucuses with 44.6% to Mike Huckabee’s 22.4% and Paul’s 17.3%. Given its electoral history and its very pronounced against the grain, independent and anti-establishment streak (it gave over 10% of the vote to Ralph Nader in 2000 and to Libertarian Ed Clark in 1980), predicting Alaska was tough. Ron Paul campaigned in Alaska, to my knowledge the only candidate to do so, and Alaska’s alleged libertarianism favoured him. Ultimately, Romney won narrowly, with 32.6% to Santorum’s 29% and Ron Paul’s rather underwhelming 24% in the state where he perhaps had the best chance of winning.
An entrance poll would have been interesting, but obviously none was taken in remote Alaska. The map of results by district gives us the next best clues about who won what in Alaska. Unlike in 2008, Romney seems to have lost the Mat-Su valley (which goes from Anchorage to Fairbanks) to Santorum. The Mat-Su is the most conservative region in Alaska and it was where insurgent candidates Steve Forbes and Pat Buchanan had done best. Romney still won a handful of districts in the Mat-Su, but Santorum likely won it overall. In Anchorage’s suburbs of sorts in the valley, Romney prevailed in Palmer but Santorum had the upper hand in Sarah Palin’s world-famous hometown of Wasilla (where she voted for Gingrich).
Romney performed best in Anchorage, where Paul also won a few precincts. He, of course, dominated with over 40% of the vote in Anchorage’s wealthier neighborhoods. He also won in Juneau, the state’s liberal state capital, and on Kodiak Island, which appears to be fairly moderate. Ron Paul’s best performance was in Fairbanks, where he apparently did best around the more liberal neighborhoods around the university while Santorum (and Gingrich, who won a district in Fairbanks) did better in the conservative areas around the military base and North Pole. Ron Paul also won the bulk of the bush. No caucuses, it seems, were held in extremely remote Bethel, Barrow and the Outer Aleutians.
The next states to vote are Kansas, which holds caucuses on March 10; and the twin primaries in Alabama and Mississippi on March 13. Romney is unlikely, at this point, to win any of these three contests, unless there is a major division of the conservative vote between Santorum and Gingrich. Mike Huckabee won the Kansas caucuses in 2008, and realistically Santorum should do very well there. Alabama and Mississippi are not as clear. Newt Gingrich could perform well in these Deep South states, and even stand a chance at winning one or both of these states. Rick Santorum, on the other hand, showed in Tennessee and Oklahoma that he has expanded his social conservative base into the South and will likely emerge with more momentum than Gingrich from Super Tuesday. The demographics of either Alabama and Mississippi are hardly receptive to Romney, given that his traditional base of seniors or affluent, educated suburbanites are not really important in either state. His only chance to win these states would be a moneybombing (and it would take a lot of money, lots of it) or hoping for a split in the conservative vote. If the results in Tennessee, Georgia, South Carolina and even Florida are any indication; then Romney should lose pretty handily in both these Deep South states.
Mitt Romney will win the nomination, but up until this point he has faced tremendous resistance from the party’s conservative base, which still hasn’t warmed up to him. In Ohio and Michigan, his victories were due to more moderate conservative voters while the most conservative voters in both those states voted in large numbers for Santorum. In the Deep South, up until this point, Romney was basically shut out everywhere outside the more moderate, older affluent suburbs of the largest cities. John McCain faced a similar problem with the conservative base in 2008, but the results we saw on Super Tuesday indicate that Romney faces an ever deeper problem. McCain had been able to win some regions of South Carolina, Georgia, Tennessee and Oklahoma which soundly rejected Romney.
Posted on March 9, 2012, in Alaska, Georgia, Idaho, Massachusetts, North Dakota, Ohio, Oklahoma, Primaries, leadership contests or internal party votes, Tennessee, U.S.A., Vermont, Virginia, Washington, Wyoming. Bookmark the permalink. 4 Comments.
The NY Times (here: http://elections.nytimes.com/2012/primaries/states/wyoming) says Wyoming’s caucuses are being held on 6th, 7th and 10th March. Is that something different?
“47% of voters were evangelical or born-again Christians, and they picked Rick by a 17-point margin (47-30) over Santorum”
Would that not be an amazing double act? ;) I gently point out
@ JD; the NYT is talking about the county conventions which will be used to allocate 12/29 WY delegates; these were precinct caucuses – just some type of straw poll which also elected delegates to these county conventions. 14 At-large delegates will be elected a state convention in April.
Pingback: USA 2012: Kansas, Islands and Dixie « World Elections