As New Census Numbers Show, the Biggest Divide Isn't North v. South Anymore—It's Metro v. Rural

On Thursday, the Census Bureau released the 2020 census data for cities, counties, and other geographic subdivisions. Just as with the state-level numbers announced in April, there were some surprises. Many large metropolitan areas grew faster over the past decade than the Bureau had previously projected, with eight of the nation's ten largest cities showing an increased growth rate compared to the 2000 to 2010 period. At the same time, most of rural America shrank in absolute as well as relative terms. A majority—52 percent—of the nation's counties actually reported a smaller raw population in 2020 than they had in 2010:

Many rural areas have been stagnant and struggling for a long time, but the distinct trajectories of big-city and small-town life have seldom been as divergent—and as connected to partisan politics—as they became over the past decade. Kathy Cramer's fieldwork in Wisconsin during the late 2000s and early 2010s found a prevailing sentiment of political alienation among rural voters, who perceived their communities as suffering a steady economic and cultural decline for which they often blamed greedy and decadent urbanites. An existing collective preference for the Republican Party among this heavily white, socially traditionalist population was soon super-charged by Donald Trump's brand of nostalgic populist nationalism, producing record Republican electoral margins in rural precincts from coast to coast in 2016.

But Trump's policies as president did not solve the long-term problems facing rural residents, or reduce the financial and cultural opportunities that migration to larger population centers can offer their children and grandchildren. The invocations of a nation fallen from past glories that resonated so strongly in small-town America inspired less enthusiasm among the residents of racially diverse, increasingly well-educated metropolitan municipalities, where even the challenges of daily life—high costs of living, insufficient housing supply, traffic congestion—often reflect the byproducts of growth and success rather than decline and decay. Prosperous suburban enclaves that once served as reliable sources of support for the mass Republican Party (such as Orange County, California; Loudoun County, Virginia; and Cobb County, Georgia) continued to shift steadily toward the blue end of the partisan spectrum in response to Trump's rise.

The fundamental geographic division in American politics has traditionally been a sectional conflict setting the North against the South. The idioms of "red states" and "blue states" caught on widely after the 2000 presidential election because they could be applied to a regional divide—blue North, red South—that was already presumed to reflect the main axis of political debate and competition. But the partisan difference between large-metro and rural residents has now become much larger than the gap between northerners and southerners.

Until 1996, the difference in presidential voting between residents of the nation's largest 20 metropolitan areas and inhabitants of rural (non-metropolitan) counties resembled the difference between the South (defined here as the eleven states of the former Confederacy plus West Virginia, Kentucky, and Oklahoma) and the North (defined as all other states from the Atlantic coast west to Minnesota, Iowa, and Missouri). Between 2000 and 2008, the urbanism gap was somewhat, though not dramatically, larger than the regional gap. By 2016 and 2020, however, the partisan difference between large metros and rural areas had become fully three times as large as the North-South difference, which had visibly narrowed (from 12 to 9 percentage points) from its 2008 peak.

Look inside practically any state in the country and you'll find blue dots corresponding to its densent and most populous urban centers, each surrounded by a sea of red rural hinterlands. The regional divide has declined since 2008 because the urban precincts of the South have grown bluer over time while the rural territories of the North have gotten redder, both shedding some of their sectional distinctiveness in the face of a consistent nationwide trend. This gave Donald Trump the ability to flip a few northern states with significant rural populations from blue to red in 2016 (such as Iowa and Wisconsin), while Joe Biden likewise outperformed previous Democratic nominees in Georgia and Texas in 2020 by winning the large metro areas of greater Atlanta, Dallas, and Houston—none of which had been carried by Barack Obama in his 2012 victory.

In a widening electoral conflict between urban and rural America, one might think that it would be clear which side holds the strategic advantage. More than twice as many voters live in one of the top 20 metro areas as reside in all rural counties combined, and the results of the new 2020 census reveal that the American population is becoming more collectively metropolitan at an accelerating rate. Even in states where they control the post-census redistricting process, Republicans will face the challenge of needing to accommodate the declining numbers of their loyal rural constituency.

But as the figure above reveals, Republicans' plight as the rural party of a increasingly non-rural nation has so far been balanced out by the fact that rural America has moved toward the GOP at a faster pace since the 1990s than urban America has shifted away. When combined with the structural biases of the electoral college and Senate in favor of rural voters, the current Republican popular coalition can easily remain fully competitive in national elections. The intensifying conflict between city and country has had a number of important consequences for how each party operates, which voters it attracts, and which states and districts it is likely to win, but it does not show any signs of ending the perennially close competition for control of the federal government that has become a distinctive characteristic of our current age.

2020 Was as Geographically Polarized as 2016, But Biden Did Just a Little Better in the Right Places

The idea of Joe Biden as a depolarizing figure, someone who aimed to transcend rather than exacerbate the nation's political divisions, was both a dominant rhetorical theme of his campaign and a major strategic premise behind his nomination. In particular, Biden was supposed to be uniquely appealing to the type of white, older, modestly-educated, socially traditionalist voter who had wandered away from the Democratic Party sometime between Barack Obama's first victory and Hillary Clinton's last defeat. A candidate who won back a significant share of this electoral bloc while receiving the energized support of the groups alienated by Trumpism—cultural liberals, metropolitan professionals, young people—would be in excellent position to gain the kind of decisive national victory that many polls suggested Biden would achieve.

In the end, that didn't happen. Whatever demographic and stylistic differences distinguished Biden from Clinton or Obama failed to change enough votes to reorder the fundamental electoral constituencies of both major parties. The geographic polarization that has defined 21st century American politics remains fully intact—but a few incremental improvements in Democratic performance turned out to be just strong enough, and well-located enough, to eke out an electoral college majority.

Most states continue to be faithfully Democratic or Republican in presidential elections. By 2016, statewide popular margins had grown to differ with the national popular vote by an average of 20 percentage points, meaning that an election in which the candidates split the national vote 50-50 would produce a typical state margin that was a 60-40 landslide for one side or the other. Biden didn't succeed in healing this particular national divide, as the average state margin remained at 20 percent in 2020:

Below the state level, geographic polarization has been primarily fueled by a widening partisan gap between large metropolitan areas turning deeper blue and rural counties turning brighter red. This divide also remained virtually unchanged between 2016 and 2020. Biden very slightly outperformed Hillary Clinton in the nation's 20 largest metro areas, winning 62 percent of the two-party vote there—the best performance since Lyndon Johnson in 1964. But he only improved on her 2016 showing in non-metropolitan counties by half a percentage point—33.7% to 33.2%—despite expectations that he would prove to be a much more appealing candidate among the whites without college degrees who dominate most rural electorates. Trump nearly matched his own record rural performance of four years before, winning a greater share of the non-metro vote than even Ronald Reagan had received in his 49-state landslide of 1984:

The main reason why Biden didn't do even better than he did within Top 20 metros was a pro-Trump national shift among non-white voters, especially Hispanics. His biggest single underperformance compared to Clinton in 2016 occurred in metro Miami, but he also lost ground in greater Los Angeles, San Jose/San Francisco, Chicago, New York, Cleveland, and Orlando. This trend helped to put Florida and Ohio out of reach for the Democratic ticket, but it was otherwise concentrated in non-competitive states where the party could well afford to lose votes.

In fact, a countervailing rise in Democratic support in 2020 among the voters of metro Atlanta, Phoenix, Philadelphia, and Detroit helped Biden narrowly carry four pivotal states that Clinton had lost, while improved Democratic performance in metro Washington, Denver, Minneapolis/St. Paul, and Boston gave Biden more electoral breathing room than Clinton had had in Virginia, Colorado, Minnesota, and New Hampshire—all virtually must-wins for a Democratic presidential nominee. So while the overall Democratic vote share in top 20 metro areas only increased very slightly between 2016 and 2020, it was more fortuitously distributed from the perspective of electoral college strategy.

As the second figure above reveals, the biggest aggregate movement toward the Democrats in 2020 occurred within metro areas outside the Top 20. These other 251 metro areas cast somewhat fewer collective votes (41% of the total two-party vote in 2020, compared to 45% for the Top 20 metros). But their lower levels of racial diversity meant that the ongoing pro-Democratic trend among college-educated whites was less likely to be numerically counteracted by racial minorities' movement towards Trump. 

Biden rarely exceeded Clinton's vote share by a dramatic amount in the smaller metros, but he consistently outperformed her by 1–3 percent across the nation (except for a few heavily Hispanic areas in Texas and California). His modest improvement in places like metro Pittsburgh and Harrisburg (Pennsylvania), Grand Rapids and Lansing (Michigan), and Milwaukee and Madison (Wisconsin) turned out to be critical elements of his precarious electoral college majority, and he was also able to carry the single Nebraska electoral vote from the metro Omaha-based 2nd congressional district that, under a plausible alternate scenario, might have given him exactly the number of electors needed to win the presidency.

It's possible that Trump had a unique appeal to certain small-town voters who might be more open to voting Democratic in future elections when he isn't on the ballot. But Democrats' retrospective interpretations of 2020 are more likely to focus on the apparent patterns of defection among urban minorities than the attrition of rural whites, which is a longer-term trend with fewer obvious implementable solutions. The continued decay among less-educated white citizens presents Democratic leaders with serious challenges in building a sustainable national electoral advantage, given this constituency's especially important geographic distribution. But if the culture wars have so alienated these voters that even Scranton Joe can't win them back, Democrats may conclude that, win or lose, the party's future inevitably lies in another direction.

In the 2020 Elections, Partisan Stability Defeats Expectations of Change

Imagine someone who set off into the Siberian wilderness right after the 2018 midterms, didn't experience any of the 2020 campaign, and returned to civilization today to see the outcome of last week's election. This person would come back to the following results: 

• A close national presidential contest decided by crucial victories in Wisconsin, Michigan, and Pennsylvania.

• A set of Senate races that almost perfectly mirrored the presidential results in each state, with the lone exception of Republican moderate Susan Collins outrunning the rest of her party's candidates in Maine.

• A mostly incumbent-friendly set of House elections, with a few of the 2018 Democratic wave's biggest upset winners—Kendra Horn of Oklahoma, Joe Cunningham of South Carolina, Xochitl Torres Small of New Mexico—unable to defend their Republican-leaning seats when facing a presidential-year electorate.

• A continuation of the ongoing pro-Democratic shift in the nation's largest metropolitan areas, pushing former "red states" Arizona and Georgia into full partisan competitiveness and reducing the size of the Republican statewide advantage in Texas.

• A partially countervailing solidification of Republican electoral strength in small towns and rural areas nearly everywhere in the country except New England, keeping key states like Wisconsin and Pennsylvania highly competitive while shifting Ohio and Iowa further toward the GOP.

If this person were a Democrat, he or she might be perfectly content with an election in which Donald Trump was defeated and Nancy Pelosi's House majority remained intact. Or maybe there would be some dismay that the popular repudiation of Trump was not strong enough to deliver Democrats an outright national landslide that handed the party clear control of the Senate. Likewise, a Republican supporter might mourn Trump's defeat—or conclude that, under the circumstances, the outcome could have been even worse. 

But there would be no reason for our returning wanderer to be surprised about any of the major results of the 2020 election. The outcome represents a doggedly consistent continuation of the basic electoral fundamentals of the last decade or more: closely-matched popular support for each party, severe and growing geographic polarization, rampant straight-ticket voting, and important Republican structural advantages in the electoral college and congressional apportionment, especially in the Senate.

Those advantages meant that Biden's victory over Donald Trump was in fact very narrow; a shift of 1 percent of the vote in Georgia, Wisconsin, and Arizona would have re-elected Trump despite Biden's comfortable lead in the national popular vote. But they also increase the magnitude of Biden's achievement: the defeat of an elected incumbent president for only the fourth time in the last century, and the ejection of a party from control of the White House after just four years in power for only the second time since 1896. And the pro-Republican tilt of our electoral institutions helps to explain why Democrats were unable to grow their margin in the House of Representatives or win a majority in the Senate, which wound up tempering some party members' delight in the election results.

That disappointment was mostly a testament to the power of pre-election polling to sway expectations. Thanks to overly rosy survey results, optimistic Democrats had visions of revolutionary success dancing in their heads: Senate victories in Montana, Kansas, and even South Carolina; breakthroughs in House races in Indiana and Missouri; even Texas turning "blue" for the first time in 44 years. This wasn't just partisan daydreaming; Republican operatives told journalists that their own private polls were as potentially devastating to their party as the publicly-released data.

There's a lesson here beyond the obvious need to re-evaluate the trustworthiness of polling methodology and forecasting models. For all the ways in which American politics seems to have entered a period of rapid and disorienting change, the partisan preferences of the electorate have only become more and more entrenched over time. Events and developments that might seem inevitably transformative, from Trump's election to the COVID epidemic to piles of small-dollar donations raised by congressional candidates from Maine to Alaska, have repeatedly proven to have only minor effects on the voting choices of citizens loyally committed to their existing partisan teams. If we simply assume that this electoral stability will prevail until proven otherwise, we are much less likely to be surprised by what we see.

The New Electoral Map Isn't Very New, But Biden's Lead Keeps the Battleground From Shrinking

We're now in the home stretch of the 2020 presidential campaign, with millions of ballots already cast via early and mail-in voting. The two candidates and their advisors are now making final decisions about where to allocate resources—ad spending, voter mobilization operations, and travel by the nominees and their running mates, spouses‚ and top surrogates—based on their current appraisals of optimal electoral college strategy.

Donald Trump and Hillary Clinton actively contested 14 states between them in 2016, collectively casting 32 percent of the nation's electoral votes. In the wake of the last election, however, most analysts expected the geographic scope of electoral competition to shrink in 2020. Trump's comfortable victories in Ohio (by 8 percentage points) and Iowa (by more than 9), along with his unexpected strength across the rest of the Midwest, seemed to signal that these two perennial battleground states were no longer pivotal and might be justifiably conceded to the Republicans in the future, while the pro-Democratic shift evident in sections of the Sun Belt from 2012 to 2016 wasn't clearly strong enough yet to push traditionally "red" bastions like Georgia and Texas into legitimate partisan competitiveness.

These expectations were all perfectly reasonable extrapolations from the 2016 results. But they rested on assumptions of another tight contest in 2020, rather than the clear and consistent Democratic lead—now flirting with double digits in the national popular vote—that actually emerged. Joe Biden's overall advantage has allowed him to remain viable in Ohio and Iowa despite their recent Republican leanings while also mounting an incursion into Georgia and Texas, both uncontested by Democratic presidential candidates since the 1990s. The only two states to drop out of the battleground category between 2016 and 2020 were Virginia and Colorado, both already swiftly moving in a Democratic direction but put altogether out of reach for the Republicans by a poor national climate for the party this year.

Contrary to previous suggestions of a shrinking battleground map, then, the presidential campaigns are once again contesting 14 states in 2020 (plus the single electoral vote awarded to the winner of Nebraska's 2nd congressional district). The replacement of Virginia and Colorado with the more populous states of Texas and Georgia means that the candidates are actually fighting over more presidential electors than last time—in fact, as the figure below demonstrates, the battleground this year is the largest in terms of electoral votes since the 2000 election:

Figure adapted from Red Fighting Blue: 

How Geography and Electoral Rules Polarize American Politics

It's often hard in the moment to separate short-term variation from long-term trends, to distinguish changes that are harbingers of the future from those that are merely temporary deviations from the norm. The assumption that 2016 foreshadowed a more permanent fading of Democratic fortunes in the Great Lakes states as the party built a new regional base in the metropolitan South and Southwest was reasonable enough, but may well turn out to be at least a bit premature. Predictions that the 2020 Democratic presidential nominee would perform better in Arizona (carried once by the party in the last 17 presidential elections) than Wisconsin (carried seven times in the past eight elections) aren't consistent with the polls so far; as of today, the FiveThirtyEight polling average shows Biden leading Trump by 8 points in Wisconsin and 4 points in Arizona. 

Polls so far indicate a fairly uniform swing between 2016 and 2020. Biden seems to be improving on Clinton's performance by a roughly comparable margin pretty much everywhere, though it wouldn't be surprising if this shift turned out to be a bit bigger in the Midwest than the Sun Belt given that he is apparently doing a better job winning back senior citizens and non-college whites from Trump than he is in outperforming Clinton's showing among young voters and racial minorities.

At the same time, it's a legitimately important development that Arizona, Texas, and Georgia are all potentially competitive states in federal elections, even if (at least for now) it requires a 10-point national popular vote margin (and a record-breaking fundraising haul) for Democrats to put all three in contention. The changing partisan alignments of a few individual states should be understood in the proper context of what is overall a historically stable electoral map. But in an era of closely-matched and highly polarized national parties, even small shifts in geographic coalitions can have big consequences for the outcome of elections and the governance of the nation.

2018 Election Recap: It Ain't Over Yet!

In this age of instant hot takes and pre-written post-mortems, it's frightfully gauche for a political analyst to wait more than a minute before weighing in on a major national event. But while unrelated professional responsibilities prevented me from updating the blog until now, perhaps I can take advantage of having had a little more time to assess the evidence before adding my voice to the chorus of electoral interpreters. After all, it's been amusing to watch the conventional wisdom evolve from "the Democrats are underperforming in the House vote" to "the Democrats did fine in the House, but have to be disappointed by the Senate and governors' races" to "the Democrats engineered a big House wave, plus they also held their own in the Senate and made key gains downballot" over the 48 hours that elapsed after the first returns arrived on Tuesday night.

Besides, the election itself isn't exactly over: plenty of ballots remain to be counted in California and Arizona, while both major statewide contests in Florida are headed to recounts that have already plunged into legal challenge amid charges of fraud and maladministration. Those of us with students too young to have consciously experienced the extended postgame in 2000 will surely welcome the opportunity to guide them through a remarkable replication played out in real time over the rest of the current academic semester. And with that, some initial observations on the results of the 2018 midterm elections—or at least the results so far—with more to follow in the coming weeks:

1. The shifts in party fortunes that resulted from this week's vote are of course important, but not enough is being made of the astounding voter turnout rate—now estimated at 48.5% of eligible citizens, which would be the highest level in a midterm election since 1966 (before the national voting age was lowered to 18) and would even approach the 51.7% of Americans who turned out for the 1996 presidential election. The opposition party is typically well-mobilized in a midterm year, and Democrats certainly succeeded in stimulating exceedingly high participation by those dissatisfied with the ruling regime. But Republicans also marched to the polls to defend a president whom many had only reluctantly supported in 2016, just as pre-election indicators of interest and engagement had suggested, and succeeded in salvaging control of the Senate and a majority of state governors and legislative chambers from the national Democratic tide.

Whether they land on the pro or con side, Americans are thinking, talking, and doing politics much more since Donald Trump was elected president of the United States. Trump's ubiquity is, in general, a massive liability for his party—given the state of the economy, the Republican House majority would surely be intact today had virtually any other 2016 presidential candidate won the office instead—but it does have its specific uses, such as a super-charged rural vote that helps to deliver midwestern and southern Senate seats and governorships into Republican hands. From today's vantage point, the turnout rate in 2020 seems likely to hit or exceed 65 percent (it was 60 percent in 2016)—which would represent the highest proportion of eligible citizens participating in a national election in more than 100 years. It sure looks like we've found a solution for the much-lamented "vanishing voter" problem of past decades; weirdly enough, though, few people these days seem to be cheering that American civic virtue has been restored to a robust state of health.

2. Trump's alienation of previous Republican supporters among the white-collar professional suburban class (especially the female members thereof) continues to leave its marks on the electoral map. Most of the gains made by Democratic House candidates were located in the nation's largest metropolitan areas: greater New York, Philadelphia, Washington, Atlanta, Miami, Detroit, Chicago, Minneapolis, Dallas, Houston, Denver, Seattle, and Los Angeles all produced at least one (and, in some cases, much more than one) R-to-D seat flip. (And in metro Boston and San Francisco, there simply are no remaining GOP-held seats for Democrats to capture.) But many of these shifts are either located in states that are already solidly blue (like New York and California) or are potentially neutralized by countervailing trends in smaller cities or towns elsewhere (as in Florida or Pennsylvania), limiting the consequences for state-level partisan alignments—which remain quite stable.

And while Democrats have reason to be encouraged by rising electoral strength in Sun Belt population centers from Georgia and Texas to Arizona and Nevada, their performance in the Midwest—while markedly better than its 2016 nadir—still stopped short of a full rebound to Obama-era levels. In fact, while the pre-election polling was for the most part impressively accurate, it consistently underestimated Republican strength in statewide races in Ohio, Indiana, Michigan, Iowa, and Missouri. The Midwest will remain the nation's biggest battleground in 2020, but it's clear that Democrats can't simply depend on Trump's New York-style brashness pushing the region's voters back in their direction. (And this observation, in turn, has associated implications for a Democratic presidential nomination contest that will soon kick into high gear.)

3. While the 2018 election was largely a referendum on the president, the identity of the individual candidates still mattered as well—as confirmed by the margin of victory in the Texas Senate race. Departing, perhaps out of necessity, from its usual practice of favoring veteran elected officials in its candidate recruitment efforts, the national Democratic Party managed to assemble a very strong assortment of "not a typical politician" congressional challengers who, for the most part, proved good fits for their districts and convinced the electorate of their qualifications for office even as they lacked long public records ripe for mining by the Republican opposition.

What we don't yet know, however, is how many of these self-styled new voices will attempt to keep their distance from older generations of Democratic leaders once they take their seats in the Capitol. There's little reason to expect a collectively demanding and persistently unruly class of House freshmen à la 1995 or 2011, but the number of Democratic candidates who promised not to support Nancy Pelosi for speaker on the campaign trail this year suggests the perceived political value that lies in maintaining public independence from the existing congressional party. Pelosi herself may be safe, at least for a while—among her other advantages, there doesn't appear at present to be a clear alternative candidate for the speakership from within the Democratic ranks—but the newly-elected members will need to be given some kind of visible accommodation once they arrive in Washington, and the question of what the post-Pelosi future looks like will hang in the air even if she successfully reclaims the speaker's gavel.

Why the 2018 Election Won't "All Come Down to Turnout"

The claim that an upcoming election "all comes down to turnout" is one of the most venerable clichés in American punditry. But it's become more popular than ever in an era when the mass public is commonly characterized as consisting of two implacable partisan tribes, equally unshakable in their preferences and aversions. If virtually every potential voter is a loyal member of either the red team or the blue team, the outcome of a national election is presumably determined simply by which party can motivate its supporters to participate at the highest rate.

Swing voters are indeed less numerous than they used to be, and the geographic polarization of the American electorate has reduced the number of states and congressional districts that are politically competitive in any given contest. But at a time when the two major parties are closely matched in national strength, the voters who remain open to persuasion continue to hold a lot of electoral power. And it's far from clear whether there will be a large enough difference in the participation rates of committed Democrats and Republicans for turnout to be the primary factor deciding the 2018 election.

To be sure, evidence is piling up that Democratic voters are unusually mobilized this year compared to the recent past. A September survey by the Pew Research Center found that 67 percent of Democratic supporters reported being "more enthusiastic than usual" about voting—a much higher rate than Pew found in either 2014 (36 percent) or 2010 (42 percent). Turnout in Democratic primary elections surged to 23 million voters in 2018, up from 14 million in 2014. And the astounding fundraising totals reported by Democratic congressional candidates, fueled by an unprecedented explosion of small-dollar contributions by individual donors, surely reflects an unusual degree of engagement among politically attentive Democratic citizens—and also ensures a series of generously-funded Democratic get-out-the-vote operations from one end of the country to the other.

But 2010 and 2014 were both unusually poor elections for the Democratic Party nationwide. Improved Democratic participation in 2018 compared to the two most recent midterms may prevent another disastrous performance, but it hardly guarantees a blue wave. And while Democrats are clearly much more engaged this year than in the recent past, Republicans are not necessarily less engaged.

According to Pew, 59 percent of Republican supporters are "more enthusiastic" about voting than usual in 2018—not far behind Democrats and at least equal to Republicans' own reported enthusiasm levels in 2014 (52 percent) and 2010 (57 percent). Among respondents to a recent Washington Post poll, 81 percent of Democrats and 79 percent of Republicans claimed to be "absolutely certain to vote" in November (surely a massive collective exaggeration of the actual turnout rate, but not one that reveals a significant difference between the parties), and an NBC-Wall Street Journal survey found that 72 percent of Democrats and 68 percent of Republicans reported "high interest" in the 2018 elections. Similarly, while the GOP didn't experience a spike in participation as large as that of the Democrats this year, turnout in Republican primaries was still higher in 2018 than 2014, growing from 15.5 million to more than 19 million votes.

Moreover, few pollsters are finding that restricting their analyses to the fraction of respondents identified as certain or likely voters (as opposed to all registered voters) produces significantly more favorable results for Democratic candidates in 2018. In fact, it's relatively good news for Democrats that they don't seem to lose ground when survey analysts use a "likely voter screen" to compensate for the projected composition of this year's electorate. Republicans normally enjoy a persistently higher turnout rate in midterm elections that diminishes or disappears in good Democratic years but seldom, if ever, transforms into an actual pro-Democratic turnout advantage. As Nate Cohn of the New York Times observes, "When Democrats hold the presidency, Republicans generally have a big midterm turnout edge . . . [and] when Republicans hold the presidency, Democrats fight back to parity."

Based on the incomplete signs so far from state election officials' reported early voting and absentee balloting totals, turnout is likely to increase across the board in 2018 from its 2014 levels. Democrats, of course, are strongly motivated this year by their deep antipathy to the Trump presidency. But Republicans don't appear to be staying home either—certainly not to the degree that Democrats did during the two Obama midterms. The polarizing figure of Trump may be inspiring elevated engagement on both sides; at minimum, it seems likely that the constant public attention commanded by the current president has resulted in Americans of all partisan persuasions thinking and talking more about politics than they did before he took office.

With Democrats and Republicans both invested in this year's election, a potential nationwide blue wave will require a non-trivial proportion of voters to shift from the GOP (or third parties) in 2016 to Democratic candidates in 2018. There are 25 Republican-held House seats that were carried by Hillary Clinton in 2016, barely more than the minimum net gain (23) needed by the Democrats to take control of the chamber, and it seems unlikely that Democrats could win enough of these seats alone to gain an overall majority. But there are also 16 Republican-held seats that Trump carried with less than 50 percent of the total popular vote, 23 additional Republican seats where Trump received between 50 and 52 percent of the total vote, and another 24 seats where Trump received 53 or 54 percent of the vote. These are the pivotal districts that hold the partisan balance of power in the House. Democrats don't need to peel off a large share of voters who previously preferred Republican candidates in order to capture majority control, but merely energizing their own habitual partisan supporters is probably insufficient to flip enough seats their way absent a modicum of successful persuasion as well.

In the Senate, the Democrats' need for a lopsided advantage among swing voters is even more evident. Five of the six most vulnerable Democratic incumbents this fall were elected in 2012 even as Mitt Romney carried their states over Barack Obama (Joe Donnelly of Indiana, Claire McCaskill of Missouri, Jon Tester of Montana, Heidi Heitkamp of North Dakota, and Joe Manchin of West Virginia); the sixth, Bill Nelson of Florida, represents a state where Obama achieved a narrow plurality win. All six states shifted further toward the Republicans in 2016; these Democrats are thus dependent upon a significant share of their constituents continuing to divide their partisan preferences, and this dependence increases further with every sign of a mobilized Republican base in 2018.

According to exit polls, self-identified independents preferred Republican candidates by margins of 56 percent to 44 percent in 2014 and 59 percent to 41 percent in 2010, making an underappreciated but critical contribution to the national GOP sweep in both elections. In 2006—the last Democratic midterm victory—independents gave Democrats an equally favorable margin of 59 percent to 41 percent. Leads for Democratic candidates in 2018 voter surveys, including a persistent advantage for the party in the national generic congressional ballot, have similarly been fueled by a steady erosion of Republican support among independents since the 2016 election. The addition of these independent votes to the revved-up Democratic "resistance" seems like a formula for electoral success in November, but many persuadable voters are not as attentive to politics as strong partisans are, and their preferences are likely to be somewhat unsettled even as Election Day swiftly approaches. With so many seats in play at all levels of government, it's still too soon to tell exactly how far the swing vote will swing.

Was the Midwestern "Red Shift" More Pro-Trump or Anti-Hillary? The Answer Matters a Lot for 2018

The election analyst Nate Cohn of the New York Times published an excellent piece today arguing that the Democrats are benefiting from an electoral battleground in 2018 that is broader than was anticipated by the post-2016 conventional wisdom. In particular, he notes, Democratic candidates appear to be doing better than expected this year in heavily white, lower-education congressional districts that voted for Donald Trump. This has allowed the party to contest many more seats than it would if the scope of electoral competition were restricted to the smaller number of Republican-held well-educated suburban districts that had shifted toward Hillary Clinton in 2016.

In general, one of the most challenging aspects of analyzing events in real time is distinguishing temporary blips from more durable trends. Popular pundits and other media figures often tend to overstate the degree to which immediate events portend long-term patterns; as I noted once in another context, "There's a temptation to assume that everything new in politics is a harbinger of the future. But lots of things are dead ends: They rise, and they go away." On the other hand, we academics are often prone to the opposite bias, hanging onto familiar theories and assumptions past the point when evidence has built up that the world has indeed changed.

When it comes to the geographic polarization of American voters, there's an unmistakable decades-long trend of divergence between (pro-Democratic) metropolitan areas and (increasingly Republican) rural areas, but also an especially sharp and unprecedented increase in this gap in the 2016 presidential election—as illustrated in this summary of partisan voting in the pivotal Midwest region taken from Chapter 6 of my book Red Fighting Blue: How Geography and Electoral Rules Polarize American Politics:

The rural Midwest has been trending Republican for a long time; Bill Clinton narrowly carried it twice in the 1990s, but Barack Obama lost this vote 53 percent to 47 percent in 2008 and 57 to 43 in 2012. In 2016, Donald Trump routed Hillary Clinton here, attracting 68 percent of the two-party rural Midwestern vote—6 points better than Ronald Reagan in his 49-state landslide 1984 reelection.

If 2016 indeed represents the "new normal," than it would make sense for analysts to take a bearish view of Democratic chances in white, small-town congressional districts in the Midwest and elsewhere this year. But if 2016 was something of an aberration, and the Trump-Clinton vote does not fully reflect the relative fundamental strength of the two parties, then the map of electoral battlegrounds opens wider, and the fortunes of congressional Democrats improve accordingly.

Midterm elections are always primarily a referendum on the president, and Trump has dominated the political scene so thoroughly since he took office that this rule of thumb is likely to be especially applicable to 2018. If the remarkable Republican strength in the rural Midwest in 2016 was primarily a reflection of Trump's personal popularity, we might expect it to carry over into 2018 unless a significant share of formerly-enthusiastic Trump supporters had become disillusioned in the interim. But if the abrupt partisan shift between 2012 and 2016 visible in the figure above was largely a reflection of Hillary Clinton's personal unpopularity with rural Midwesterners—as well as a Clinton campaign that eschewed economic issues to an unprecedented degree for a modern Democrat—we shouldn't be surprised by a significant Democratic rebound in the region this November, since Clinton will be neither on the ballot nor in the White House.

Cohn's piece focuses exclusively on the House of Representatives, but the question of whether the 2016 "red shift" across the north-central section of the country is a temporary or enduring development becomes even more critical when we turn to the Senate—where Democrats are defending nine seats in Trump-carried states stretching from Pennsylvania to Montana plus two more in Minnesota (which Trump lost by less than two points). Using the 2016 presidential results as a starting point for expectations of 2018 outcomes paints a very optimistic picture for Republicans; Trump carried Indiana, Missouri, and Montana by about 20 points and won North Dakota and West Virginia by more than 35. Yet all of these states elected Democratic senators only six years ago, and all of them but North Dakota were actively contested at the presidential level as recently as 2004 or 2008. 

Many loyal Democrats will not easily accept the belief that the results in 2016 reflected a widespread popular antipathy to Hillary Clinton. And the surprising nature of Trump's victory has encouraged the view, even among his fiercest critics, that he maintains under-appreciated political strengths. But the more validity to the conclusion that Trump was a weak candidate who won a close and fluky election only because he was facing a seriously flawed opponent, the rosier the outlook becomes for Democrats this November.

Corporations and the Culture War: Op-Ed in the New York Times

I have a new op-ed piece in the New York Times explaining how the culture war has fueled geographic divisions in the American electorate, which in turn has ended up furthering the interests of wealthy individuals and corporations. The analysis draws on the research presented in my recent book Red Fighting Blue: How Geography and Electoral Rules Polarize American Politics.

In Today's America, Small Voting Shifts Can Have Really Big Effects

My new book Red Fighting Blue: How Geography and Electoral Rules Polarize American Politics demonstrates that American elections have become more geographically polarized over the past 25 years and investigates the causes and consequences of this important development. I argue that it's very difficult to understand how politics operates today without recognizing the ways in which the spatial distribution of party votes across geographic lines interacts with the rules and mechanisms of our electoral institutions to produce representative democracy in the United States.

One of the most distinctive attributes of our current political era is that the two parties are closely balanced at the national level, yet each side maintains a formidable record of electoral dominance across large regional subsections of the United States. Since the 1990s, we have experienced a series of elections in which the overall outcome has been decided by a narrow margin even as most states and congressional districts are securely and reliably either Republican or Democratic. While it was once common for most states—and nearly all large states—to be electorally up for grabs and thus actively contested by both sides in a national campaign, the number of "battleground" states has dwindled in the 21st century to a small minority of the nation:

At the same time, the election-to-election stability of these state alignments has reached historically unmatched levels. It's not just the case that most states in any given year will be safely pro-Democratic or pro-Republican while a shrinking remainder of states will be open to contestation—instead, it's increasingly the same states that are predictably "red" or "blue" in election after election,  and the same few swing states, like Ohio, Florida, and Michigan, that hold the national balance of power from one election to the next. Over the past five presidential elections, 37 states and the District of Columbia have aligned with the same party in each contest; as the graph below shows, this represents a record degree of state-level consistency. We're a long way from the presidential elections of 1956 and 1964, when all but five states in the nation voted for opposite parties in two races held just eight years apart.

Because of the winner-take-all American electoral system, these trends mean that partisan shifts among a relatively small subset of voters in a relatively small section of the nation can have a tremendous effect on outcomes if they happen to be concentrated in exactly the right place. My political science colleague Seth Masket wrote today in Vox about the changing partisan preferences among non-college-educated whites in the small-town Midwest that proved electorally pivotal in 2016, delivering the White House to Donald Trump despite his loss in the national popular vote (see my own graph below as well). Seth concludes, I think persuasively, that these shifts are not broad enough to add up to a party "realignment"—but they still proved to be decisive because they occurred in a few key states that were otherwise closely divided between the parties. Hillary Clinton ran particularly well among college-educated whites and racial minorities in 2016, which helped her outperform previous Democratic candidates in non-battleground states like California, Texas, and Arizona but didn't pay her any dividends in the unique arithmetic of the electoral college.

The Geographic Polarization of the Midwest, 1980–2016

Because the leadership of the two parties is as mutually polarized and otherwise differentiated as at any time in modern history, the consequences of these electoral outcomes for the trajectory of national politics are increasingly significant even as the outcomes themselves are more and more likely to be contingent on the attributes, and eccentricities, of the electoral system itself. Historically speaking, there wasn't much of a difference between how Americans voted in 2012 and how they voted in 2016. But there's an awfully big difference between President Barack Obama and President Donald Trump.

New Book: Red Fighting Blue

I'm very happy to announce that my new book, Red Fighting Blue: How Geography and Electoral Rules Polarize American Politics, has just been published by Cambridge University Press in paperback, hardcover, and eBook formats.

The book explains how the emergence of significant and persistent regional differences in partisan voting patterns in both presidential and congressional elections since the 1990s has had a profound effect on party politics in the United States. The American electoral system, with its geographically-defined voting constituencies and winner-take-all rules, has greatly magnified these differences when translating the preferences of citizens into electoral college outcomes and the partisan and ideological composition of Congress. The moderate Democrats formerly elected in large numbers from what are now the "red" states and the moderate Republicans who once represented the "blue" states are both disappearing from office, leaving increasingly polarized parties and an ever-shrinking scope of electoral competition. In an age in which the two major parties are closely balanced but each increasingly dominant across large regional subsections of the nation, the specific ways in which the geographic distribution of party support interacts with the rules of the American electoral process accounts for how our politics works—or doesn't work—in the 21st century.

Three Reasons Why There Are So Many Republican Governors

A recent analysis by Matthew Yglesias notes that the current Democratic hold on the White House masks much deeper partisan weakness at the congressional and, especially, state level. Republicans currently control both the governorship and state legislature in fully half (25) of the states, while just 7 states (California, Connecticut, Delaware, Hawaii, Oregon, Rhode Island, and Vermont) remain under unified Democratic rule. (In the remaining states, power is divided between the parties.) Yglesias concludes that the Democrats are in "deep trouble" as a party and need to formulate a strategic plan to regain electoral strength in non-presidential contests. In response, Phil Klinkner argues that such trends simply reflect the historical tendency of the president's party to suffer significant losses in midterm elections; if and when Democrats lose the presidency, they will likely make compensatory gains at the state level.

There is much to say about this topic, but I will focus today on the partisan control of state governorships, where Republicans hold a pronounced advantage. There are currently 31 Republican governors, just 18 Democrats, and one independent (Bill Walker of Alaska, who was elected with Democratic support). Three main reasons explain why 62 percent of the states are now governed by members of the GOP:

1. There are more red states than blue states, and especially more deep red states than deep blue states. The Cook Partisan Voting Index, a measure of state partisanship that denotes how many more percentage points a state voted in either the Democratic or Republican direction than the national outcome in the last two presidential elections, reveals that 21 states are more Republican than the nation as a whole by a margin of more than 5 points; 13 of these states are more than 10 points redder than the national electorate. By comparison, just 12 states have a Cook PVI of more than 5 points in the Democratic direction, only 4 of which (Hawaii, New York, Vermont, and Rhode Island) are more than 10 points bluer than the nation itself.

Between the 1960s and the 1990s, states often departed from their normal partisan alignments in presidential elections when choosing governors and other state officials—especially in the South, where substantial Democratic strength in state politics endured long after the region shifted decisively toward the GOP in presidential contests. But American voters are now highly likely to support the candidates of a single party across multiple elective offices, and the growing partisan polarization of state-level electoral outcomes has produced a flotilla of securely red states stretching across the South and interior West. For example, Republican voters in states like Nebraska, Wyoming, and Kansas once crossed party lines to elect Democratic governors with some regularity, but this tendency has become rare in the strongly partisan contemporary electoral climate.

2. Most purple states swung to the Democrats in 2008 and 2012, but to the Republicans in 2010 and 2014. The nation's most politically competitive states, almost all of which narrowly supported Obama in both of his presidential victories, largely shifted in the opposite partisan direction when their own governorships were up for election two years later, in a manner consistent with Klinkner's description of a thermostatic reaction against the presidential party. (The weak national economy that benefited Obama in 2008 also surely contributed to the 2010 losses of numerous incumbent state-level Democrats who had been swept into office by the anti-Bush wave of 2006.) In the 2010 midterms, Republicans defeated or succeeded sitting Democratic governors in the purple states of Ohio, Michigan, Wisconsin, New Mexico, Pennsylvania, and Iowa while narrowly retaining the Florida governorship as well; except for Pennsylvania's Tom Corbett, who was damaged politically by accusations of receiving improper gifts and by his maladroit handling of the Penn State football scandal, all of these new Republican governors won second terms in 2014—buoyed in part by the same economic recovery that had assisted Obama's reelection two years before.

3. Republicans found good electoral fortune in blue states. Sometimes sheer luck simply favors one party or the other at any given time. For example, Democrats achieved a filibuster-proof Senate majority in 2009, which was ultimately critical to the passage of the Affordable Care Act, by exploiting what later turned out to be a tainted criminal conviction secured by prosecutorial misconduct in Alaska and by winning a statewide recount in Minnesota by a margin of 0.01 percent. More recently, Republicans have enjoyed the opportunity to run against an unusually high number of unpopular Democratic incumbents or weak Democratic challengers in traditionally blue states, including Martha Coakley in Massachusetts, Anthony Brown in Maryland, Jon Corzine in New Jersey, and Pat Quinn in Illinois. In addition, left-leaning Maine has twice elected Tea Party Republican Paul LePage to the governorship by a popular plurality due to the presence of a spoiler candidate who split the Democratic vote in both 2010 and 2014.

***

Reasons 2 and 3 for the Republican gubernatorial advantage almost certainly reflect temporary factors that will swing back in the Democrats' favor sooner or later. Popular Republican incumbents in purple states will be termed out of office, Democratic candidates will improve in quality (or it will be the Republicans' turn to squander a series of winnable races by selecting especially unattractive nominees, as they may have done this year in Kentucky and Louisiana), and an eventual Republican presidential administration will provoke a midterm backlash that will benefit the Democratic opposition. It is the first factor—the preponderance of red states over blue states—that represents the most serious long-term obstacle to majority Democratic control of state governorships. The political "reddening" of the South and interior West over the past 25 years is primarily due to the prevailing cultural conservatism of these regions' inhabitants, which has increasingly bolstered popular support for the Republican Party up and down the ballot. The suggestion that national Democratic leaders do what they can to remedy this serious electoral vulnerability is no doubt sensible; assuming that they are unlikely to renounce their party's increasingly liberal positions on social issues, however, it may be hard to find a simple solution to the Democrats' red-state blues.