Heading into the 2016 presidential election cycle, the most influential guide for political journalists was a 2008 book called The Party Decides. Written by four eminent political scientists, it explained that for several decades presidential nominees have effectively been chosen by unelected political insiders, as candidates fight in “invisible primaries” for endorsements by prominent politicians and interest groups. The voters, it argued, tended to ratify these choices and rally around candidates with widespread and prestigious support.
But like John Kenneth Galbraith’s 1967 book The New Industrial State, which argued that big corporations, tempered by big government and big labor unions, determined the course of the economy, The Party Decides turned out to be a better description of the recent past than an accurate forecast of the near-term future. Political science, despite its name, is not a science, and generalizations about presidential elections are risky because there have been so few of them—only 46 since something like the current two-party system sprang into existence in 1832 and only 11 since primaries started dominating the selection of party nominees in 1972. When I was in the political polling business, I was told not to base conclusions on the responses of subgroups comprised of fewer than fifty respondents. Scholars of presidential elections, even if they go back to the days when Andrew Jackson faced off against Henry Clay, have less data to work with than that.
Certainly few analysts in May 2015, just 14 months before the national conventions, predicted that Donald Trump and Bernie Sanders would be serious competitors for the Republican and Democratic nominations. Neither the New York real-estate billionaire nor the Vermont socialist had significant support from his party’s elected officeholders or party officials. Indeed, each had received endorsements from only a handful of party insiders up through the conclusion of the primary season 13 months later. Yet Trump won 42 percent of the votes cast in Republican primaries and caucuses up through the Indiana primary on May 3, 2016, after which his remaining two opponents withdrew. By the last contest on June 7, he had won 44 percent. Sanders had won 43 percent of votes cast in Democratic primaries and caucuses, but that was not enough to defeat Hillary Clinton in what was effectively a two-candidate contest. But Trump’s vote totals were enough to secure a delegate majority in a race that had started off with 17 serious candidates. “The party” got the Democratic nominee of its choice, but only after a longer struggle and by a narrower margin than it surely imagined, while “the party” was utterly foiled in the Republican contest despite an impressive array of attractive and competent candidates.
So why has this presidential campaign cycle been different from all other presidential campaign cycles? And is the general-election campaign likely to be as different from other general-election campaigns as the primary contests were different from their predecessors?
One way to look at this election is as a collision of an irresistible force with an immovable object. This irresistible force is the widespread discontent with the direction of the nation today. The immoveble object is the persistent partisan divisions that have prevailed and intensified in presidential, congressional, and state elections over the past twenty years.
The sources of the irresistible force of discontent are not hard to discern. After resurgent growth and victory in the Cold War in the 1980s, and continuing economic growth in the 1990s, the 21st century brought Americans 15 years of mostly sluggish growth and a series of mostly unsuccessful, or at least inconclusive, foreign military interventions. Major legislation passed by one-party votes, notably the 2009 stimulus package and the 2010 Affordable Care Act, have proved to be far less popular than their sponsors expected. Major bipartisan legislation, frequent in Bill Clinton’s presidency and the first term of George W. Bush’s, has become rare if not extinct, with a President lacking the inclination and skill to negotiate and a Republican House majority often unwilling to trust its leadership.
This discontent found an outlet in the disruptive candidacies of Donald Trump and Bernie Sanders. Each attracted constituencies different from those in his party’s recent nomination contests. Republicans in 2008 and 2012 were divided between countryside and suburbs, between white Evangelical Christians and less intensely religious groups. The divisions can be seen in the critical contests between John McCain and Mike Huckabee in 2008 and between Mitt Romney and Rick Santorum in 2012. In both cases the eventual nominee piled up big majorities in the relatively affluent and somewhat less Evangelical suburbs, while his opponent carried rural areas and small towns, but not by enough votes to prevail.
In 2016 the divisions were different. White evangelicals did not vote solidly for any candidate, but split their votes between Trump, Ted Cruz, and Marco Rubio. Large suburban counties in many states gave Trump pluralities or even majorities. One clear pattern is that Trump ran better among voters without college degrees (“I love the poorly educated!” he exclaimed after winning the Nevada caucuses) than college graduates, but he got sizeable numbers of votes from graduates as well. Certain demographic groups resisted Trump’s appeal: Mormons, Dutch-Americans in northwest and central Iowa and western Michigan, German- and Scandinavian-Americans in Wisconsin and the Upper Midwest states. Other ethnic groups tilted toward Trump. A majority of Italian-Americans live within a hundred miles of New York City, and in that arc Trump won more than 50 percent of the votes, including 81 percent in heavily Italian-American Staten Island. In addition, he ran strongest not in Florida’s Southern-accented congressional districts, but in those with the largest number of migrants from New York and the Northeast. Examining the returns, I argued that Trump fared poorly with those groups with large degrees of what scholars Charles Murray and Robert Putnam have called social connectedness or social capital, and did very well with groups with low social connectedness. His percentages in Appalachia—from southwest Pennsylvania through Tennessee, northern Alabama, and Mississippi—were especially large.
The Democratic primaries saw some reversals of the Party’s trends in 2008. That year, Barack Obama beat Hillary Clinton among black voters by wide margins. In 2016 Clinton won black voters over Bernie Sanders by similarly large margins in the South and somewhat smaller margins in the North. However, in 2008 Clinton dominated Appalachia; in 2016 it voted for Sanders. In both elections Clinton tended to carry Hispanic voters, but in 2016 she did significantly less well among white voters without college degrees, and Sanders tended to carry small towns and rural counties not only in his home area of New England but across the Midwest and in white-majority regions in the South.
Both nominees owed their victories to large majorities cast by their parties’ most downscale constituencies—blacks, especially Southern blacks, in the Democratic Party, whites without college degrees in the Republican Party. For the Democrats, the result conformed to the thesis of The Party Decides. Hillary Clinton had the lion’s share of endorsements from party officials and office-holders, but she won because of big margins from black voters; exit polls suggest she did no better than break even among white Democratic voters. For the Republicans, the result was directly contrary to the wishes of almost all party insiders. Trump benefited greatly from his celebrity, gained as a publicity-hungry real estate mogul even before he became a reality-television host, and he dominated news coverage of the Republican race—getting more airtime on cable and broadcast news than the 16 other candidates combined from the day in June 2015 when he rode the escalator down to the Trump Tower lobby to announce his candidacy.
Trump benefited as well from the dynamic of a multi-candidate race, in which it is not in the interest of one candidate to attack another: an attack by candidate A on candidate B may hurt B, but it is likely to hurt A as well and help C or D or E. That is exactly what happened when Chris Christie launched an attack on Marco Rubio in the debate days before the New Hampshire primary. Rubio lost his chance for a second-place finish that might have propelled him to a one-on-one contest with Trump, while Christie ran poorly and left the race the day after the primary. The New Hampshire results encouraged Jeb Bush to stay in the race through the South Carolina primary, where his percentage added to Rubio’s was just 2 percent lower than Trump’s, and it gave life to the candidacy of John Kasich—who, unlike Rubio, carried his home state on March 15 and remained in the race until May, though he only carried seven counties outside Ohio. Trump’s best-organized rival, Ted Cruz, did win the Iowa caucuses, but lost by agonizingly narrow margins in multi-candidate races in Arkansas, Kentucky, Louisiana, and Missouri. He won a solid victory over Trump in heavily German-American and high-social capital Wisconsin on April 5, but lost in demographically diverse Indiana on May 3 and exited the race.
Trump’s victory was not inevitable, but it was perhaps overdetermined. He was an outsider candidate, defying political correctness, in a year when discontent was an irresistible force. Not considered a standard politician, he announced his candidacy when the Republicans’ leading contender was the son and brother of two former Presidents and the Democrats’ was the wife of another. He criticized both parties while wooing a Republican electorate not only harshly critical of Barack Obama, but also of many of the programs of George W. Bush and of the current Republican congressional leadership.
So the irresistible force of discontent prevailed in the race for the Republican nomination and came closer to prevailing than almost anyone expected in the race for the Democratic nomination.
But how will it fare against the immovable object of persistent strong partisan attachments in the general election? The most downscale constituencies of each party have produced two nominees who are unpopular with a majority of general-election voters—how will it play out in November?
The partisan deadlock has been as protracted any in our history. Democrats have won four of the last six presidential elections, and a plurality of the popular vote in a fifth, while Republicans have won majorities in the House of Representatives in nine of the last 11 congressional elections. But neither side has won by big margins. No major-party nominee has won less than 46 percent of the popular vote or more than 53 percent over the past 16 years, or the past 24 years if you allocate Ross Perot voters in 1992 and 1996 to their second-choice candidates. This is in vivid contrast to the four decades after World War II, in which incumbent Presidents of both parties in times of perceived peace and prosperity won re-election by landslide margins in 1956, 1964, 1972, and 1984. In those years most voters remembered the horrors of the Depression and World War II and were glad to cross party lines for Presidents who seemed to produce better times.
Today most voters remember what seemed to be the more peaceful and prosperous decades of the 1980s and 1990s, a time in which party preferences were shaped increasingly by cultural issues. In the 1990s the demographic variable most highly correlated with voting behavior was religion, or degree of religiosity, with the most observant in each sectarian group veering Republican and the less observant or secular veering Democratic. As a result, suburban voters in major metropolitan areas outside the South moved toward the Democratic Party, while there was a countervailing but quantitatively smaller shift toward the Republican Party outside major metropolitan areas. At the same time, partisan preference in presidential and congressional elections converged. From 1968 to 1988 Republicans won five of six presidential elections, by an average margin of 10 percent of the popular vote, while Democrats won majorities with at least 243 seats in the House. Starting in 1992 Democrats have won most presidential elections, but by an average margin of only 4 percent of the popular vote, and Republicans have won most congressional elections but failed to hold more than 242 House seats until they won 247 in 2014.
This partisan deadlock has resulted in an unusually stable electoral map by historical standards. Only three states changed their electoral-college votes between 2000 and 2004; only two did so between 2008 and 2012. The list of 11 target states has become familiar even to those who are not political junkies, and campaigns have concentrated most of their organizational and advertising efforts there. Voters have responded accordingly. Total voter turnout sagged between 2008 and 2012, but it was up 0.8 percent in the 11 target states, while it fell 2.7 percent in the rest of the nation. It has become easy to predict how three-quarters or more of the states will vote in presidential elections, even as it has remained difficult to predict which candidate will win.
This immovable object may prove movable in November 2016, though probably not as movable as suggested by the predictions in some quarters that Donald Trump would lose by a wide margin, with something like the 38 percent won by Barry Goldwater and George McGovern in the postwar elections of 1964 and 1972. Pre-national convention polling indicates that, despite the turmoil of the primary campaign season, solid majorities of self-identified Republicans and Democrats are prepared to vote for their parties’ nominees. Neither Donald Trump nor Hillary Clinton has received 50 percent in just about any national poll, and polls in the 2012 target states—which are less frequent and perhaps less reliable—have seldom shown either with overwhelming leads.
But there are some discernible differences from previous electoral cycles. Trump is polling as well as or better than Mitt Romney did in 2012 among whites without college degrees but is running perceptibly weaker among white graduates—results in line with the divisions in the Republican primaries. Clinton is running far ahead among blacks and Hispanics. But it’s unlikely that she will be able to equal the turnout or Democratic percentages among blacks achieved by the first black President. And Trump’s poll numbers among Hispanics—a more varied group, with quite different partisan leanings in different states—are roughly similar to what Romney was polling four years before. Young voters seem highly hostile to Trump, but young women as well as young men voted heavily for Sanders and against Clinton in the Democratic primaries, and their support for Clinton over Trump seems less than enthusiastic.
This suggests that Trump may be highly competitive in target states with older and less-educated populations, such as Florida, Pennsylvania, and Ohio. Together they have 67 electoral votes, without which Barack Obama would not have been re-elected in 2012. It also suggests that Trump’s prospects are significantly less favorable in target states with younger and more-educated populations, like Colorado, Virginia, and North Carolina. These have 37 electoral votes, with North Carolina’s 15 going to Mitt Romney in 2012.
And the list of target states could conceivably be expanded. Pre-election polling has shown Trump weaker than previous Republican nominees in heavily Republican states and Clinton somewhat weaker than previous Democrats in some heavily Democratic states. Democrats’ hopes of carrying Arizona and Georgia, with their large Hispanic and black populations, could be realized if Trump fails to match previous nominees’ large share of white votes. Polls have even shown a close race in Utah, Romney’s strongest state, whose Mormon majority has shown a clear distaste for Trump. Some supporters of Trump have claimed, unconvincingly, that he could win heavily Democratic New York and New Jersey and, more convincingly, that his anti-free-trade positions could enable him to carry Michigan and Minnesota. Both have been absent from previous target-state lists, but voted only 54 and 53 percent for Obama in 2012.
Moreover, there appear to be more undecided voters than in recent elections. At this stage in 2008 and 2012, fewer than 10 percent of respondents to most polls said they were not voting for either major-party candidate. In 2016 pre-convention polling that percentage has been higher, around 15 percent. When respondents are given a choice of voting for Libertarian candidate Gary Johnson and Green Party candidate Jill Stein, the number who choose neither major-party candidate rises to about 23 percent. Polls giving respondents a choice of third- and fourth-party candidates probably overstate respondents’ actual support for these candidates, which has typically decreased or evaporated when it comes time to actually vote. With sizeable majorities of voters expressing unfavorable feelings toward both Clinton and Trump, it’s plausible that larger than usual percentages are unwilling to commit to either and could change their minds during the course of the campaign.
And it’s plausible as well that many of these people will simply not vote. Voter turnout surged enormously during the Bush presidency, from 105 million in 2000 to 131 million in 2008. But contrary to popular impression, it has sagged perceptibly during the Obama presidency, in both the 2012 presidential election (from 131 million to 129 million) and the 2014 House contests (from 86 million in 2010 to 79 million). In a nation closely divided between two partisan blocs, differences in turnout can produce differences in results. Democrats hope that voters antagonized by Trump will turn out in large numbers, but there is little history in presidential elections of high turnout motivated by negative feelings toward a candidate. Trump backers hope that voters energized by Trump’s unorthodox messages will turn out in great numbers, noting that Republican primary turnout in 2016 surged far ahead of 2008 and 2012 levels while Democratic primary turnout lagged behind that of 2008. But the evidence suggests that Trump’s specific appeal was responsible for less than half the increased primary turnout.
Both parties face difficulty in maximizing turnout for their sides. Trump won the Republican nomination without any large organization and in the weeks running up to the national conventions did little to assemble one, apparently intending to rely on Republican National Committee efforts. The Clinton campaign and the Democratic Party appeared to have a much more effective turnout apparatus, and one that delivered victory in 2012. But even then Barack Obama received 3.5 million fewer votes than he had four years before.
The Clinton camp is likely to have difficulty matching the 2008 Obama campaign’s success in mobilizing support from young voters. The exit poll that year showed Obama leading John McCain by 66 to 32 percent among those under thirty; his popular-vote margin among that age group amounted to 7 percent of the total electorate, identical to his overall popular-vote margin. In the 2014 House election, in contrast, the Democratic margin among young voters was 54 to 43 percent and, with lower turnout, amounted to only 1.5 percent of the total electorate. Young people tend to move frequently, have few community ties, and be less interested in politics and government than older people. Clinton’s weak showing among young voters in the primaries suggests they lack enthusiasm for her despite their evident distaste for Trump.
The Trump camp is likely to have difficulty maximizing turnout as well. His strongest support in primaries came from those with low social connectedness, who are presumably hard to contact and mobilize. For all his success in the primaries, he had won just 42 percent of Republican primary and caucus voters when he clinched the nomination in the Indiana primary on May 3.
So it remains an open question how the seemingly irresistible force of public discontent will shift the seemingly immovable object of partisan deadlock. There have been many surprises in the 2016 presidential election cycle so far. There may be many more ahead. The days of party decisions may be over. But it’s not clear who will be deciding now.