Thursday, May 07, 2015
Earlier this week, we debuted our initial Crystal Ball Electoral College ratings in Politico Magazine. We’ve reprinted that column below for those who did not see it. As promised, we have elaborated on the map and our reasoning for the initial judgments.
In our internal deliberations on these initial ratings we quickly agreed on a large majority of the ratings, which shouldn’t come as a surprise given the rigidity of the modern Electoral College.
We had some differences of opinion about the Leans Democratic states in the Midwest and the Northeast, as well as a handful of states that typically go Republican. For instance, we considered starting Pennsylvania and Wisconsin as Toss-ups to reflect how close they could be in a tight national election. However, given that both states have voted Democratic even in years when the Democratic nominee has lost (2000 and 2004), and because neither state has shown a clear pro-Republican trend in recent presidential elections, we could not justify portraying either state as a coin-flip to start.
We also debated what to do with Arizona, Georgia, and Missouri, states that have consistently voted Republican presidentially since 2000 yet have yielded very close results (Missouri) or have demographic trends favoring the Democrats (Arizona and Georgia). They are certainly not Safe R -- or not the way they used to be, at least in the case of the latter two -- but if the GOP nominee is losing any of them, he is almost certainly on his way to a large national loss. Therefore, Likely R is the logical place for them.
Indiana was easier. Barack Obama’s 2008 win was something of a fluke; it was only the second time since the end of the World War II that the state voted Democratic, and Mitt Romney strongly restored it to the GOP column in 2012. So we’ve started it as Safe R.
Finally, after a vicious argument that resulted in bloodletting, New Hampshire was designated Toss-up instead of Leans Democratic, for reasons we explain more fully below.
As Map 2 shows, the Electoral College has been very stable over the past four presidential elections. Remarkably, 40 of 50 states have backed the same party from 2000 to 2012, a whopping 80%. Of the other 10 states, five have gone with one party 75% of the time and the other five have split their results evenly between the two parties in that period. We’ve discussed how much more rigid this map appears compared to previous periods.
Note: Doesn’t include Washington, DC, which supported the Democrats in each election from 2000 to 2012 -- and of course has backed the Democratic nominee without exception since it first cast electoral votes in 1964. (One DC elector refused to vote for Al Gore in 2000, but she chose no other candidate either.)
Yet there is another quartet of elections in U.S. history that displayed similar geographic inflexibility. From 1876 to 1888 -- four bitterly contested elections -- 30 of 38 the states that existed at that time (79%) backed one party in all four elections, as shown in Map 3. Six others supported the same party in three of four elections, while just two were 50%/50% states. And the number of 100% same-party states would be higher if not for the 1876 race, when Florida, Louisiana, and South Carolina backed Republican Rutherford B. Hayes in that disputed election; with the end of Reconstruction, Democrats regained hegemony in those final ex-Confederate states to produce what became known as the “Solid South.”
The Sunshine and Buckeye states are, respectively, the third and seventh largest in the Union, and they have a combined 47 electoral votes. Both states were superfluous in President Obama’s two victories: Take away either state (or both), and Obama still wins. But George W. Bush would not have won in either 2000 or 2004 if he had failed to carry these states.
While there have been exceptions, such as Ohio in 2004, these states have typically leaned a bit more to the Republicans at the presidential level. Practically speaking, the GOP doesn’t have a path to victory without both states. If a Republican can’t win Ohio, can he win other competitive Midwestern states that are usually more Democratic, such as Pennsylvania or Wisconsin? That’s very doubtful as the Keystone and Badger states have been more Democratic than Ohio in every presidential election going back to 1964, when Lyndon Johnson performed slightly better in Ohio than Wisconsin. Pennsylvania has been more Democratic than Ohio in every election since 1948.
Without Ohio and probably some more electoral votes from the Midwest, the GOP path to victory is difficult to imagine.
Similarly, if Republicans can’t win Florida, they would likely have to make it up in the Midwest and Pennsylvania. Here’s what that map might look like, which we computed on 270toWin, where one can spend hours playing with the Electoral College.
Is this map possible? Sure. But it’s not plausible for 2016. It would require the Democratic nominee to win Florida while losing six states where Obama did better in 2012: Colorado, Iowa, New Hampshire, Ohio, Pennsylvania, and Wisconsin. If whites trend even more disproportionately toward Republicans over time, this theoretically could be a winning GOP map in the future, but a growing Democratic minority vote across the country might well cancel out any Republican gains with whites.
In 2012, Romney crushed Obama among white voters nationally, 59%-39%, but lost by almost four percentage points and 5 million votes. A debate has raged over whether the GOP should concentrate its efforts on making inroads among nonwhite voters, who comprise an expanding portion of the electorate, or instead work to increase turnout among the majority portion of the white vote that prefers Republicans.
We recently explored Obama’s approval rating at the national and state level and found some inconsistencies, notably in the two whitest swing states, Iowa and New Hampshire. These also happen to be the only two states where Obama won the white vote in 2012. The president’s approval numbers in these states have been lower than we might expect based on the 2012 results. Could the GOP be in a position to improve on Romney’s performance among whites in the battleground states? Table 1 presents the exit poll data on the 2012 white vote in the 10 states we view as Toss-ups or Leans for 2016.
In these 10 states, Romney won the white vote in eight, yet Obama claimed nine of them overall, including the three whitest: Iowa, New Hampshire, and Wisconsin. Should Republican performance among white voters in these states improve, the GOP nominee would be in a good position to win Iowa and New Hampshire (given that their electorates are more than 90% white). Meanwhile, stronger Republican performance among whites could also open up the possibility of winning Colorado, Nevada, Ohio, and Wisconsin. Considering how well Romney did, it may be a taller order for the 2016 GOP nominee to perform better among whites in Florida, North Carolina, and Virginia.
Connected to the questions about the white vote are the possible new paths to an Electoral College majority for Republicans. One such route may wind through the Midwest. Map 5 shows the lean of the states based on their 2012 result in relation to the national margin. For instance, Obama won nationally by 3.9 points and won Ohio by 3.0 points. So Ohio was actually 0.9 points more Republican relative to the national outcome in 2012.
In Table 2, the 10 states that were either less than five points more Democratic or less than five points more Republican, relative to the national outcome, are listed. Of those states, four are in the Midwest, all won by Obama in 2012: Iowa, Minnesota, Ohio, and Wisconsin. Romney won 206 electoral votes in 2012, and this quartet holds 44 that could be added onto the Republican total under the right circumstances.
However, Minnesota is a real stretch for the GOP. The Land of 10,000 Lakes last went Republican in a presidential election in Richard Nixon’s 1972 landslide victory. In Ronald Reagan’s crushing 1984 win, Minnesota native son Walter Mondale won just his home state (and Washington, DC). We have Minnesota rated Likely Democratic along with Michigan, another Midwestern state that always tantalizes Republicans but never delivers on Election Day.
After removing Minnesota from the equation, Iowa, Ohio, and Wisconsin are left and they add up to 34 electoral votes. Potentially, they would give Republicans 240 EVs, with Romney’s 2012 performance as the base. Florida’s 29 electoral votes could then get the Republican to 269 electoral votes, which is technically a tie but would almost certainly end up as a Republican win, for reasons we explained earlier this year (the Republican-dominated House would decide the election).
So adding Iowa and Wisconsin to Ohio and Florida, while holding everything else constant from 2012, would produce a GOP victory. However, winning Wisconsin will be difficult, even with Gov. Scott Walker (R-WI) on the Republican ticket. The state last voted for a GOP presidential nominee in 1984, though George W. Bush did come within less than half a point of winning it in both 2000 and 2004, so victory there is far from impossible.
If anyone suggests that Democrats have a shot at Texas in 2016, or Republicans can capture California, Illinois, or New York, ignore them. Texas simply is not diversifying at a rate anywhere close to what Democrats would need in order to seriously contest the state, and the giant Democratic vote in New York City, Chicago, Los Angeles, and San Francisco locks in the three pillars of the Democrats’ big-state dominance for the foreseeable future. If any of these states become competitive, there is far bigger story brewing, mainly that an expectedly close national election has turned into a broad win for one side or the other.
Democrats in places across the Appalachian region -- a sprawling collection of 428 counties and independent cities (in Virginia) that runs from Western New York all the way down to northern Alabama and Mississippi -- are hopeful that Hillary Clinton will revive their fortunes in a region that intensely dislikes President Obama. There is some merit to this argument if one looks back to the 2008 Democratic primary: Clinton won 374 of the 428 counties/independent cities that make up Appalachia, including every county in West Virginia (the only state that is 100% Appalachian, as defined by the Appalachian Regional Commission), every Appalachian county in Ohio, and all but two of Pennsylvania’s 52 Appalachian counties (the exceptions being Centre County, home of Penn State University, and neighboring Union County).
But the problem for Clinton and the Democrats, post-Obama, is that their decline in the region goes far beyond presidential results. Democrats have lost at least a dozen U.S. House seats in Appalachia in the Obama years. Once long-time Rep. Nick Rahall (D, WV-3) failed in his reelection bid last year, the Democrats said goodbye to their last truly Appalachian foothold. In 2014, Sen. Shelley Moore Capito (R-WV) became the first Republican to win a Senate seat in the Mountain State since the 1950s, and she swept every county. Incumbent Democratic senators such as Pennsylvania’s Bob Casey and Virginia’s Mark Warner, who had shown prior strength in the region, saw many of those counties turn red in their 2012 and 2014 reelection bids, respectively (Warner barely hung on).
Thus, even though Bill Clinton twice won states like Kentucky and West Virginia, there’s no reason to expect Hilary Clinton can win or really even compete in either. The same goes for Arkansas, the state where she served as first lady when her husband was governor for a dozen years. While Arkansas is not Appalachian, it is similar to Kentucky and West Virginia in that it has a largely white population that is culturally conservative and has made a sharp turn toward Republicans in recent years. Combined, Republicans now control 12 of 13 House seats in all three states; Democrats hold just one seat, based in Louisville.
However, there is one way in which a bit of an Appalachian bounce-back could help Democrats. If Hillary Clinton (or another Democratic nominee) can attract more votes out of Appalachian Southeast Ohio or Western Pennsylvania than Barack Obama did -- which is not impossible to imagine -- it could help make up for slightly lower black turnout in places like Cleveland and Philadelphia.
We wondered whether we should rate states conditionally based on the strengths or weaknesses of individual candidates who might win the GOP nomination. For instance, is Wisconsin a Toss-up as opposed to Leans Democratic if Scott Walker is the nominee? Would Jeb Bush or Marco Rubio move Florida to Leans Republican if one of them is the nominee?
The answer, at least at this early point, is no.
In the case of Walker, the state is so divided politically, and Walker is such a polarizing figure, that we don’t believe he would have much of a home field advantage. And Walker is not especially popular right now: The state’s gold standard survey, the Marquette Law School Poll, recently found his job approval at just 41%. Walker, despite winning three statewide elections in four years, has never dealt with a presidential-level electorate. The average voter turnout in Walker’s midterm victories in 2010 and 2014 and recall win in 2012 was about 55%. But presidential turnout in 2012 was about 73%, an 18-point difference. The electorate Walker would face in 2016 will be far bigger and more Democratic than any he has seen in his statewide elections.
The same is true for Bush and Rubio in Florida. Neither has faced a presidential-sized electorate, and this fact looms large for Bush because he hasn’t been on the ballot since 2002, while Rubio has only appeared once on a statewide ballot.
Now, it’s possible that if one of these men won the nomination, polling might confirm a home field advantage of sorts. The last two presidents did get something of a boost in their home states. In 2004 Illinois was 6.5 points more Democratic than the nation as a whole, but then zoomed to nine points more Democratic when home-state Senator Obama won in 2008. (Illinois dipped back down to 6.5 points more Democratic in 2012.) In 1996, Texas was eight points more Republican than the nation. That increased to 11.5 points more Republican when Gov. George W. Bush got elected in 2000, and it was 10.5 points more Republican when Bush won again in 2004. But these states were uncompetitive, and it’s also impossible to confidently attribute all of the variance in the results just to a home-state effect. Florida and Wisconsin will be hotly contested in 2016, and the flood of advertising and organizing could serve to blunt any home-state edge while reinforcing partisan divisions.
It’s worth mentioning that Gov. John Kasich of Ohio, a potential Republican candidate, has strong approval ratings back home (61%, according to Quinnipiac University), although he is a less plausible nominee than Bush, Rubio, or Walker, and the same partisan forces that would apply to any of those other candidates would probably also apply to him if he were the nominee. But he might have the best case of any of the potential nominees for president for prompting a ratings change in his home state, though as with the others we would need strong polling data to back that up.
We have ordered our Toss-up and Leans states in Table 3. You may notice that Pennsylvania, which we rate as Leans Democratic, was not as Democratic as some of our Toss-up states and could credibly be paired with the Midwest path to victory discussion for the GOP in point five. Meanwhile, New Hampshire, a Toss-up, was actually slightly more Democratic than Pennsylvania in 2012: Obama won 51.98% in the former and 51.96% in the latter.
So what gives? Nate Silver of FiveThirtyEight examined the concept of electorate elasticity prior to the 2012 election. New Hampshire has one of the most elastic electorates according to his analysis, and the 2012 exit poll data helps us see why. The Granite State electorate had easily the largest percentage of self-identified independents (or something else) of all the battleground states, 43%. While most independent identifiers are closet partisans, the exit poll data suggest that, compared to other battlegrounds, New Hampshire may really have a larger block of voters who are potentially persuadable.
Note: Percentages may not add up to 100% because of rounding in the exit poll.
Another reason for this is that most of the Granite State electorate is white. Although white voters backed Romney by a 20-point margin in 2012 nationally, this paled in comparison to nonwhite voters, who supported Obama at a more-than-80% clip. So New Hampshire’s relative absence of diversity means that it lacks a large group of nonwhite voters who can be viewed as relatively reliable Democratic votes, particularly African Americans.
This is a key difference with Pennsylvania. Of our 10 battleground states, the Keystone State had the largest percentage of Democratic self-identifiers in 2012, in part because its electorate was 13% black and 21% nonwhite as a whole. At the same time, Pennsylvania also had the largest percentage of Republican self-identifiers, meaning that the state had by far the lowest percentage of independent identifiers in these states, 20%. The next-closest? North Carolina with 29%. These figures show why the Pennsylvania electorate is actually relatively inelastic because there are fewer potential swing voters and more base voters who will support their side in most every circumstance.
One other major difference between small New Hampshire and big Pennsylvania: In all likelihood, the Democratic nominee is going to get at least a 400,000 vote-edge out of Philadelphia County, and that might be low (Obama got close to a 500,000-vote lead there in 2012). It’s hard for Republicans to make that up in the rest of the state.
Thus, despite New Hampshire and Pennsylvania having nearly identical election results in 2012, the Granite State is a Toss-up because its electorate is among the most elastic, making it less predictable, whereas the Keystone State is Leans Democratic because its inelastic electorate gives Democrats a high floor.
An easy reason Virginia might be the key swing state: In 2008 and 2012, Virginia’s presidential result was closest to the national outcome. Virginia’s 52.63% for Obama in 2008 was 0.23 points less than Obama’s national percentage of 52.86%, and its 2012 percentage of 51.16% for Obama was 0.15 points greater than Obama’s 51.01% nationally. While traditional big-time swing states Florida and Ohio are slightly more Republican than the nation as a whole, Virginia has been precisely where the country has landed the last two presidential cycles.
However, 2012 marked the first time since 1944, FDR’s final reelection, that Virginia’s result was more Democratic than the nation’s as a whole. On top of this, Virginia Democrats won every statewide election in 2012, 2013, and 2014, though Sen. Mark Warner had an extremely close call that few saw coming in 2014’s low-turnout midterm. Nonetheless, Warner won despite a poor environment that saw Democrats lose Senate seats in swing states like Colorado and Iowa, the former having another Democratic incumbent who entered office at the same time as Warner (now ex-Sen. Mark Udall). Gov. Terry McAuliffe’s (D) 2013 victory broke a run dating back to 1977 of the president’s party losing a gubernatorial race in Virginia. The “presidential curse” was a somewhat coincidental streak that would eventually be broken; what’s important is that, on the whole, Democrats have been very successful in recent elections in the Commonwealth. The demographic shifts and massive growth in Northern Virginia have been almost the sole driver of the state’s move to the left, though improved Democratic performance in the entire “Urban Crescent” -- Northern Virginia, Hampton Roads, and Greater Richmond -- has made Democrats the stronger party in recent statewide contests.
Should the Democratic nominee manage to win Virginia again -- remember, McAuliffe is one of the closest Clinton allies -- it’s hard to see Republicans winning back the White House in 2016. Our map starts out with the Democrats favored for 247 electoral votes, so capturing Virginia plus Colorado and Nevada, both of which were more Democratic-leaning in 2008 and 2012 than the Old Dominion, would put the Democratic nominee over the 270 mark in the Electoral College.
If the Republican just barely wins, we think the map is likeliest to look like this:
If the Democrat just barely wins, we think the map is likeliest to look like this:
What do these two maps tell us? Basically, that if a party wins both Colorado and Virginia, it will win the White House. Both states have hewn closely to the national electoral average for the past two presidential elections, and they are formerly reliable Republican states that have moved in a Democratic direction. The chances of one of them providing the decisive 270th electoral vote to the winner, as Colorado did in 2012 (with Virginia right behind as just a shade more Republican), are high.
Larry J. Sabato is the director of the Center for Politics at the University of Virginia.
Kyle Kondik is a Political Analyst at the Center for Politics at the University of Virginia.
Geoffrey Skelley is the Associate Editor at the Center for Politics at the University of Virginia.
See Other Commentary by Larry Sabato
See Other Political Commentary by Kyle Kondik
See Other Political Commentary by Geoffrey Skelley
See Other Political Commentary
Rasmussen Reports is a media company specializing in the collection, publication and distribution of public opinion information.
We conduct public opinion polls on a variety of topics to inform our audience on events in the news and other topics of interest. To ensure editorial control and independence, we pay for the polls ourselves and generate revenue through the sale of subscriptions, sponsorships, and advertising. Nightly polling on politics, business and lifestyle topics provides the content to update the Rasmussen Reports web site many times each day. If it's in the news, it's in our polls. Additionally, the data drives a daily update newsletter and various media outlets across the country.
Some information, including the Rasmussen Reports daily Presidential Tracking Poll and commentaries are available for free to the general public. Subscriptions are available for $4.95 a month or 34.95 a year that provide subscribers with exclusive access to more than 20 stories per week on upcoming elections, consumer confidence, and issues that affect us all. For those who are really into the numbers, Platinum Members can review demographic crosstabs and a full history of our data.
To learn more about our methodology, click here.