One Election Night a year ago didn’t get us to this place.

Long before Donald Trump upset Hillary Clinton for the presidency last year, the nation was facing steadily rising social and political tensions rooted in diverging reactions to the relentless economic, demographic and cultural changes reshaping American life. But the bruising 2016 race between Trump and Clinton widened these divisions to a new extreme. In the starkly contrasting pattern of support each candidate inspired, the election functioned something like a bolt of lighting on a starless night: It illuminated, with sudden starkness, a political landscape deeply fractured along lines of race, generation, class and geography.

One tumultuous year later, and one year before the 2018 midterm election, those fissures look only more imposing.

Far from seeking to bridge these divides, Trump, as both candidate and president, has repeatedly demonstrated he believes it benefits him to widen them. Trump’s willingness, even eagerness, to push at the most volatile fault lines in American life, from race to religion to gender, has created an explosive new reality. The underlying changes remaking America are so disruptive that this would be a tense period no matter who held the highest positions of political leadership. But confronting those changes with a President whose bottomless appetite for both cultural confrontation and personal feuds adds layers of volatility.

Trump is something of an arsonist: He seems to take delight in burning down rather than building up.

“What is qualitatively different about Trump than anybody else who came before him is he’s willing, and seems to delight, in lighting the tinder and creating a conflagration, whereas others tried at least now and then to dampen the tinder and to keep the conflagration from happening,” says Peter Wehner, the director of strategic initiatives in the George W. Bush White House. “No one did it perfectly and of course, within a certain range, politicians will get into fights that are to their advantage. What’s really different is Trump is something of an arsonist: He seems to take delight in burning down rather than building up.”

The intensity of emotion Trump has inspired, among supporters and opponents alike, has unleashed destabilizing pressures in both parties and opened a gulf between the places where he is revered and reviled. Trump has riveted many voters in the parts of America, primarily outside of the largest cities, that feel most eclipsed by growing racial and religious diversity and the evolution toward a more globalized and post-industrial economy that relies less on fossil fuels. But he has outraged, even terrified, many voters in the mostly urbanized parts of America that welcome all of those trends.

Each side appears increasingly uncomprehending of the other – and increasingly dubious that it’s possible, or even desirable, to bridge their differences. It is into this pool of combustible tensions that Trump, on issues from whether National Football League players should kneel for the national anthem to whether the Charlottesville protests contained “very fine people” on each side, is routinely dropping matches. “The combination of these underlying factors and Trump’s emotional and psychological state,” says Wehner, “is pretty explosive.” The explosions Trump detonates almost daily have thrust Washington into perpetual turmoil -- and promise the same for the 2018 midterm elections that will offer voters their first broad chance to render a verdict on his tumultuous presidency.

The fuel Trump plays with has been gathering for some time.

At the core of America’s modern political divide is a convergence of propulsive changes in demography, culture and the economy. “There’s a lot of things hitting the country at once,” says Robert P. Jones, president of the non-partisan Public Religion Research Institute.

Demographically, the nation is living through the most profound transformation since the Melting Pot era at the turn of the 20th century. Almost 40% of the total population is now non-white, roughly double the share in 1980. Among the young, the change is even more accelerated. Kids of color represent about half of all Americans 10 and younger, and since 2014, they have constituted a majority of all K-12 public school students nationwide. William Frey, a demographer at the Brookings Institution, has calculated that from 2000 to 2014, not only did whites decline as a share of the under 18 population in 46 of the 50 states – but so did the absolute number of white kids. “The 2020 census is going to show that the under 18 population is majority minority, same as the under 10 population now is, and that there is an absolute decline of white youth in the US,” Frey predicts flatly.

A majority of public school students in the United States are now nonwhite

The percentage of white public school students has shrunk from 61% at the turn of the century, and is projected to fall to just 45% in the next decade.

Closely related to the nation’s growing diversity is the increasing prominence of immigrants. People born abroad now constitute about 14% of all Americans. That’s the highest total since the years around World War I and nearly triple the 5% level in 1965, when Congress replaced the restrictive laws from the 1920s that had severely limited immigration for four decades. Mark Hugo Lopez, director of Hispanic research at the Pew Research Institute, projects that under current law, first-generation immigrants will exceed 15% of the population by some time around 2025, breaking the previous record high reached in 1890.

The Share of 1st or 2nd generation immigrants in the United States

More than a quarter of Americans are 1st or 2nd generation immigrants, the highest levels in the United States since 1940.

More racial diversity has contributed to another tectonic shift: Increasing religious pluralism. For almost all of American history, people who identified as both white and Christian represented a majority of the American population. Through the 1960s, about eight in 10 Americans identified as white Christians. That number had declined to only slightly less than seven-in-10 by the time of Ronald Reagan’s reelection in 1984 and still stood at nearly two-in-three when Bill Clinton won his second term in 1996.

But the steady increase in the non-white population, and a steady decline in the share of Americans who identify with any Christian faith, pushed white Christians below half of the population for the first time around 2012, according to surveys of religious preference by Pew and others. That erosion has continued unabated since: an extensive PRRI poll recently found that white Christians had fallen to just 43% of the population. Non-white Christians account for just over one-fifth of the population while Americans unaffiliated with any religious faith now represent nearly one-in-four.

White Christians no longer make up majority of American people

The share of white Christians has fallen to just four in 10 over the last decade, with dramatic drops across evangelical protestants, mainline protestants and Catholics.

Partly because of these shifts in religious allegiance, the nation has experienced a rapid change in cultural mores. Fifteen years ago, same sex marriage was not legal in any state and faced opposition in polls from a significant majority of Americans; now it is legal everywhere, with support from a significant majority. Mixed race marriages have grown more common. Debates over the rights of transgender people have opened a new frontier in the cultural conversation.

And even as these demographic and cultural shifts have rolled through American life, the economy has undergone an equally wrenching restructuring. In 1965, the core blue-collar industries of manufacturing, construction and mining (including energy production) accounted for over one of every three American jobs. Now that number is less than one-in-seven and the Bureau of Labor Statistics projects it will fall below one-in-eight in by 2024. Job growth today is driven much more by post-industrial occupations, like health care, education, business services and tourism. Blue-collar jobs accounted for just three of the 30 professions that the BLS recently projected would grow the fastest through 2026. Growth is also concentrating more into large metropolitan areas that are racing into the information economy and integrating into the globalized market for products, people and ideas.

Over roughly the past two decades, attitudes toward these enormous changes have become the fundamental dividing line in American politics. In both presidential and congressional races, Republicans rely on what I’ve called a “coalition of restoration” that revolves around older, blue-collar, and evangelical Christian whites, mostly outside of urban areas, who feel most uneasy about these changes. Democrats mobilize a competing “coalition of transformation” centered on minority, millennial and college-educated white voters (especially women), who are mostly clustered in major metropolitan areas and the most comfortable with the changes.

More explicitly than any other recent Republican nominee, Trump ran as a candidate of restoration. His backward-facing promise to “make America great again” soldered a powerful connection with all those who feel eclipsed by these changes. With her competing message of “stronger together,” Clinton arguably sought even more than Obama did to portray herself as the champion of the transforming urbanized America.

Trump paved a path to an electoral college victory by consolidating his core groups of older, blue-collar, and evangelical whites just slightly more than Clinton consolidated her core groups of minorities, millennials, and well-educated white women. But the brutal competition left the two sides representing coalitions that diverged far more, both geographically and demographically, than the parties did in earlier generations.

The religious divide between the parties offers one powerful example. In the new PRRI study, about three-fourths of Republicans still identify as white Christians, comparable to the nation overall in 1984. In stark contrast, only about three-in-10 Democrats now identify as white Christians. Nearly as many Democrats are unaffiliated with any religion, and just over one-third are non-white Christians. Those two groups, though, each represent only about one-in-every-nine Republicans.

The divide extended to less obvious distinctions. Trump won 26 of the 30 states where the foreign born represent the smallest share of the population. Clinton won 16 of the 20 states where they represent the highest share.

Trump won 13 of the 16 states that produce the most natural gas, 11 of the 15 that produce the most coal, and 16 of the 20 that produce the most oil. In a cumulative measure of reliance of manufacturing and resource extraction, Trump won 27 of the 32 states, almost entirely across the nation’s interior, with the highest per capita levels of the carbon dioxide emissions linked to global climate change. Clinton won 15 of the 18 states with the lowest per capita carbon emissions – most of them largely post-industrial states along the two coasts.

In all these ways and more, the two parties now glare at each other across the divide of what America has been, and what it is becoming.

It looks like one party that is holding on to a 1950s America’s demographics and increasingly looks like a white Christian party that is going to be perpetually tempted toward nationalist parties around that identity. And then we have a Democratic Party that is following these (demographic and economic changes, and might, on the other hand, be tempted to double down on (pursuing) everyone but white Christians.

“It looks like one party that is holding on to a 1950s America’s demographics and increasingly looks like a white Christian party that is going to be perpetually tempted toward nationalist parties around that identity,” says Jones, author of the 2016 book The End of White Christian America. “And then we have a Democratic Party that is following these (demographic and economic) changes, and might, on the other hand, be tempted to double down on (pursuing) everyone but white Christians. In a country with a two party system that is a pretty volatile mix: Race, religion and identity overlaid with partisanship.”

Just how volatile that mix can be has been explosively evident since Trump took office. Trump’s presidency has offered a precarious balancing of the unprecedented and the conventional. In style, he has been unlike any previous president, precipitating an unending succession of feuds with a rotating cast of foils and antagonists; regularly delivering false accusations that are easily disproven; and attacking the legitimacy of any institution, from the media to the courts, that he believes can resist him. Parts of his agenda have been equally unconventional, as he’s embraced policies on both immigration and trade far more insular than the GOP has endorsed before.

And yet in other ways, he has proven a much more conventional Republican than he signaled during the campaign. He’s pushed elements of the traditional evangelical social agenda on limiting access to contraception in health care or rolling back protections for transgender soldiers more enthusiastically than almost anyone expected. And after a campaign where he promised to defend government programs that could benefit his older and blue-collar white base, he’s emphatically endorsed the traditional modern Republican goals of cutting taxes, spending and regulation.

Despite all the attacks on Republican establishment and particular Republican leaders, for the most part he’s really allowed congressional Republicans to define his agenda, especially the details of his agenda.

“Despite all the attacks on Republican establishment and particular Republican leaders, for the most part he’s really allowed congressional Republicans to define his agenda, especially the details of his agenda,” notes Emory University political scientist Alan Abramowitz.

Indeed, amid all the daily tumult, an informal division of responsibility within the GOP appears to be emerging. The Republican-controlled Congress and executive branch departments, like the Environmental Protection Agency, are systematically advancing the traditional GOP goals of cutting taxes, spending and regulation and funding the Pentagon. Meanwhile, Trump is provoking a procession of twitter-fueled confrontations primarily around cultural issues, many of them (from Charlottesville to the NFL to immigration) with a sharp racial edge. “This is something he is comfortable doing and it has become ritualized,” says Alan Wolfe, a retired political scientist at Boston College who has extensively studied America’s divisions. “It’s like a play and everybody knows the plot.”

The policy agenda “is not something he cares very deeply about and he’s not active in it,” adds Wehner, now a senior fellow at the conservative Ethics and Public Policy Center. “There’s no energy behind it. Most of his energy seems to be directed on cultural divisions, on racial divisions and on creating a more fractious and fractured society.”

Though the strains are evident, this informal division of responsibility appears to be preventing a full-scale break between Trump and congressional Republicans, despite his frequent sniping at them. It’s less clear the GOP electoral coalition can withstand the strain.

Many Trump supporters remain convinced that he is doing exactly what they sent him to Washington to do. But compared to his vote in the 2016 election, Trump has lost support in office from both sides of his coalition: His approval rating today in surveys is consistently lower than his share of the vote last November among both blue-collar and older whites (especially women) and white-collar whites.

Yet Trump’s difficulties with upscale whites look like a greater long-term risk to him and the GOP. The danger that well-educated whites, who usually tilt Republican, will recoil from Trump’s definition of the party has been symbolized by the procession of business leaders who abandoned White House advisory councils after his widely criticized response to Charlottesville, and the unprecedented criticism he has absorbed from the three previous GOP presidential nominees (Mitt Romney, John McCain and George W. Bush), each of whom has accused him of promoting racial divisions at home and/or undermining America’s historic international role.

“I think the white working class vote is more loyal to him…and pulling off of him is harder even if the job approval drops,” says veteran Democratic pollster Stanley B. Greenberg. “Whereas the college educated voters, absent Hillary (Clinton) on the ballot, have a lot of room to cast an anti-Trump vote.”

The biggest question looming over the 2018 midterms is how these complex currents of opinion about Trump will affect the contest. In many respects, 2018 is shaping up as a classic collision between an irresistible force and an immovable object.

The irresistible force is the widespread discontent over Trump’s first year. Throughout American history, the president’s party has almost always lost House and Senate seats in the election two years after he first takes office. Indeed, the last three times a president entered a mid-term election with unified control of Congress -- Bill Clinton in 1994, George W. Bush in 2006 and Barack Obama in 2010 -- voters have revoked it.

Several other signals point to 2018 risk for Republicans. One is a potential intensity gap. Though Trump inspires passionate support, the share of Americans who say they strongly disapprove of his performance is consistently much larger than the share who strongly approve. And Democrats now consistently hold an average lead of about eight percentage points when voters are asked the “generic” question of which party they intend to support in the 2018 elections.

These early advantages in public opinion may prove ephemeral, but they have produced tangible benefits for Democrats. Troubles for a new president almost always encourages more retirements from House members in his party, and Republicans have faced a trickle of high-profile retirements (such as Charlie Dent, Ileana Ros-Lehtinen, Dave Reichert and Dave Trott) that insiders fear could become a flood later this year. As is often the case when a president from the other party stumbles, Democrats have also found it easier than usual to recruit strong candidates. That’s been the case not only in the 23 GOP-held congressional districts that voted for Clinton last fall, but also in a number of places that Trump carried. And the antipathy to Trump has helped several Democratic challengers unexpectedly outpace their Republican opponents in early fund-raising.

With all of these gales blowing, even some leading Republican thinkers are bracing for a blustery mid-term. One is Tom Davis, the former Republican congressman from Northern Virginia and chair of the National Republican Congressional Committee, who is one of the party’s shrewdest (and most encyclopedic) strategists.

I felt the day after the 2016 election that now that the Republicans controlled everything they were going to have a tough midterm. I still continue to feel that and I see nothing to change that.

“I felt the day after the 2016 election that now that the Republicans controlled everything they were going to have a tough midterm,” says Davis, who is now the director of federal government affairs for Deloitte. “I still continue to feel that and I see nothing to change that. If you look at some of these state House seats (around the country) that flipped (from Republican to Democrat) in special elections this year, you are seeing what you would expect -- that Democrats are juiced and Republicans are dispirited. And that’s a bad formula going into a midterm.”

Yet all of these seemingly irresistible forces threatening the GOP face an immovable object: The structural advantages Republicans enjoy in the battle for control of Congress, particularly in mid-term elections like 2018.

These start with the electoral map. Analysts across the ideological spectrum agree that Democrats need to win more than half of the total votes in House races to win a majority in the chamber. That’s partly because Democratic voters are overly concentrated in major urban centers, but mostly because Republican control of state governments after the 2010 census allowed them to gerrymander districts that favor them. Some analysts believe Democrats need to win the House popular vote by as much as eight points (roughly their current lead in the generic ballot test) to capture a majority, though others, like Abramowitz, put the number slightly lower. Whoever is right, that imbalance means that to recapture the House, Democrats will likely need to win at least some seats where Trump is more popular-potentially much more popular-than he is nationally.

Democrats face a similar structural disadvantage in the Senate because the Constitution’s allocation of two senators for every state magnifies the impact of less populated, predominantly white and culturally conservative rural states that firmly favor the GOP. This year, the imbalance is especially pronounced. Democrats are defending ten Senate seats in states that voted for Trump in 2016, mostly across the industrial and agricultural heartland, while Republicans are defending seats in only two states where Clinton ran well: Nevada (which she won) and Arizona (which she narrowly lost).

Republicans today are most optimistic about their prospects against Democratic Sens. Claire McCaskill in Missouri, Joe Donnelly in Indiana and Bill Nelson in Florida (if GOP Governor Rick Scott runs). But Democrats could also face competitive races in Wisconsin, Ohio, North Dakota, Montana, Michigan, Pennsylvania and West Virginia, all heavily blue-collar states that Trump carried. Meanwhile, beyond Arizona and Nevada, the only state that now seemingly provides Democrats any chance to win the third seat they would need to recapture a Senate majority is Tennessee, if former Democratic Governor Phil Bredesen joins the race to replace retiring Republican Bob Corker. And even that would be a steep climb given the state’s underlying partisan direction.

Beyond the map, the usual composition of the midterm electorate also hurts Democrats. The party’s modern alliance of minorities, millennials, and college-educated whites is a boom and bust coalition because the first two groups are less likely to vote in the midterms than presidential elections.

Greenberg recently sounded an alarm that, despite Trump’s unpopularity, Democratic-leaning voters appear no more inclined to vote in 2018 than those who lean toward the GOP. Given Trump’s strong connection with his base, Democratic strategist Jesse Ferguson says the party must plan for the possibility the GOP will not suffer the turnout slump that usually afflicts the president’s party in mid-term campaigns.

“If 2006 was energized Democrats and depressed Republicans and 2010 was energized Republicans and depressed Democrats, I’m pretty convinced 2018 will have energized Democrats, but I’m not convinced yet that it will have depressed Republicans,” Ferguson said. “So we may be in for energized vs. energized and I don’t know what that creates.”

This complex ledger of factors has divided election prognosticators. Those who take the micro approach -- analyzing races district-by-district, candidate-by-candidate -- are generally skeptical Democrats can win the 24 Republican seats they need to recapture the House. Democrats, for instance, are focusing intently on seven Republican held House seats in California that voted for Clinton over Trump last year. But Darry Sragow, publisher of the non-partisan California Target Book, points out that in five of those seven seats, Democrats other than Clinton have won no other races, such as state legislative contests, since the districts were drawn in 2012. “They are not easy pickings,” says Sragow, a former Democratic strategist. “Based on the historic performance of those districts, they are not likely to win most of them.” In all, the respected Charlie Cook Political Report still identifies 228 seats as either safe, likely or leaning toward the GOP.

On the other side are those who focus on the macro factors in elections. They note that wave elections often sweep out incumbents who objectively had no business losing -- and they see signs that Trump’s unpopularity could be generating such a wave against Republicans. “It feels to me you may well have the potential for that kind of reaction against Trump,” says Greenberg, the Democratic pollster.

Those plotting the party’s official strategy at the Democratic Congressional Campaign Committee are more cautious: They say swing voters still see Trump as such a singular and unconventional figure that it’s not clear they will punish House Republicans even if they have soured on the President. For that reason, Democratic candidates, in their early sparring, are mostly focusing more on linking their Republican rivals to the unpopular GOP congressional leaders than to Trump.

And yet whether Democrats buy ads tying Republican House members to Trump or not, and whether those Republicans embrace Trump or keep their distance, history suggests that attitudes toward him will cast a huge, potentially decisive, shadow over next year’s congressional election. According to media exit polls, at least 82% of voters who say they approve of the President’s performance have voted for his party’s House candidates in every midterm election since 1994 (except in 1998, when 77% did.) Over that same period, at least 82% of voters who disapproved of his performance have voted for the other party. Politicians can run from a president of their own party -- but they can’t really hide any longer.

That dynamic underscores what may be the safest prediction for 2018. Whether or not Democrats win the 24 seats they need to recapture the House, or, less likely, find the three they need to retake the Senate, the election seems probable to further the parties’ demographic and geographic separation.

In the House, the Democrats’ best prospects are the 23 GOP members in districts that Clinton carried over Trump; about three-fourths of those seats are crowded with the kind of college-educated professionals uneasy about the president. Even in a bad overall environment, Republicans in turn could post some further gains in the 12 Democratic-held House seats that Trump carried -- most of them largely blue-collar districts outside of urban areas. Similarly, the Senate results could continue the Republican advance in predominantly white, blue-collar and older Rustbelt states (like Missouri and Indiana) while marking further Democratic progress in diversifying and younger Sunbelt states including Nevada and Arizona.

In that way, the 2018 election could further the partition of America into distinct spheres of influence. And that would only intensify the polarization swirling around a tumultuous president whose actions -- and the counter-reactions they provoke -- harden that separation every day.

“These things reinforce each other,” says Wolfe. “Trump’s actions will cause more protests from African-Americans and minorities, which will then fuel Trump. It is a hard, negative vicious cycle to break. I think we are going to long for the old culture war, because you could find some kind of compromises there. Everybody said you could never resolve abortion. But we in fact did. I don’t see how you reach a compromise about this level of anger and resentment.”

Correction: This story has been updated to accurately reflect the percentage of white Christians in the United States.