By 1990 it was widely held by political journalists that Republicans had a “lock” on the presidency and Democrats a “lock” on the House of Representatives. Pundits and political scientists identified structural advantages that Republicans had in presidential elections and Democrats in congressional elections—advantages they were confident would persist for years and perhaps decades to come. As it turned out, they didn’t: just when political trends and their causes are identified, they often tend to disappear. I remember speaking to a group of House Democrats in 1990 and arguing that these “locks” might get picked in the coming decade, and so the 1990s might see the election of President Bill Bradley—I, or maybe the country, picked the wrong Bill—and Speaker Newt Gingrich. At that point a groan came from the vicinity of John Dingell, who was on the verge of becoming the longest-serving member of Congress and ended up maintaining that status longer than anyone else in history. Since my talk, Democrats have won four of the seven presidential elections starting in 1992 and Republicans have won majorities in the House of Representatives in ten of thirteen congressional elections starting in 1994. The old “locks” are history.

The first lock to be picked was the Republicans’ supposed lock on the presidency. The idea was that with Republicans now sweeping the South and always carrying California (as they had, with the single exception of 1964, between 1952 and 1988), they were assured of something like 215 electoral votes, and were far ahead in almost all the Great Plains and Rocky Mountain states’ 70 electoral votes. This analysis was not airtight. Some southern states, economically and demographically, were coming to resemble northern states open to voting Democratic, and Republicans carried California only by single digits in the close elections of 1960, 1968, and 1976, and in the not-so-close election of 1988.

Nevertheless, Democrats who were raised to consider themselves the natural majority party were nevertheless enormously frustrated at losing five of the six presidential elections from 1968 to 1988, and by an average of 10 percent of the popular vote. In the three presidential elections of the 1980s, they carried a total of only 17 states out of 150 cumulatively up for grabs. Despite their continued majorities in the House of Representatives and most state legislatures during this period, this protracted failure at the national level eventually prompted rethinking the national party’s stands. The shrewdest effort came from Bill Clinton, governor of Arkansas, and the Democratic Leadership Council founded by Al From, a former congressional staffer. They recognized that Democrats’ stands on crime and welfare dependency were responsible for losing almost every crime-ridden metropolitan area, and that Democrats’ dovish statements on foreign policy were unsustainable politically as the United States won the Cold War and the Soviet Union ceased to exist. In early 1991 the Democrats seemed destined to lose a fourth election in a row. After the American victory in the Gulf War—which Senate Democrats, still in the grip of Vietnam, almost unanimously opposed—George Bush’s job approval peaked at 91 percent, and prominent Democrats like Mario Cuomo, the governor of New York, declined to run. Bill Clinton entered the race that fall, while tacitly admitting his frequent philandering.

But the unexpected happened. In February 1992, the Texas celebrity billionaire Ross Perot launched an independent candidacy, bemoaning Bush’s NAFTA trade agreement with Mexico and Canada. In the words of a Democratic strategist, Perot “de-partisanized the critique of Bush” in a way that Clinton, struggling through the gauntlet of Democratic primaries, could not. In June, Perot led polls in a three-way race, with Clinton running third. But in July, Perot dropped out of the race in the third day of the Democratic National Convention and Clinton’s support rose 25 points. The Republican leads that appeared so steady in four of the five elections of the last 25 years now seemed to vanish. Perot reentered the race in October, and despite his weird conduct he won 9 percent of the popular vote; he carried no states but ran second in Maine and Utah. Bush, who had won 53 percent four years before (more than anyone since then), got only 37 percent this time—a decline in percentage comparable to that of Herbert Hoover during the Great Depression. Clinton won 43 percent and a solid 370 electoral votes, including 253 from 22 states Bush had carried in 1988. But the electorate was very much split: Clinton got more than 50 percent only in his own state of Arkansas and the District of Columbia; Bush and Perot ran under 50 percent everywhere.

In many ways the Perot candidacy resembled the Trump candidacy 24 years later. Both were widely known as successful entrepreneurs; both decried the free–trade policies of both major parties and questioned the worth of foreign alliances and military interventions. Both campaigned bombastically and disregarded traditional political customs. Both zoomed to leads in the polls, Perot against both major party nominees within four months of announcing, Trump against sixteen Republican rivals within two months of announcing. But Perot stumbled, and his defeat may have shown Trump the unwisdom of running an independent candidacy, however tenuous his ties to either major party, and certainly demonstrated the unwisdom of abandoning his supporters by dropping out of the race and then hopping back in.

The patterns of partisan support in the 1992 election were significantly different from those that had prevailed in most of the country since 1952 and in the South except for those years with a southern Democratic or third–party candidate. Clinton’s overall percentage mostly trailed that of Michael Dukakis in the two-way contest four years earlier, but the articulate Clinton ran significantly better in both upscale and blue–collar suburbs in the largest metropolitan areas and urban conglomerations—New York, Chicago, Los Angeles and San Francisco, South Florida. Clinton’s promise to end “welfare as we know it” and his firm disapproval of violent crime fared much better in those high-crime, high-welfare areas than Dukakis’s tendency to see welfare recipients and criminals as victims of racism or inequality. At the same time, Clinton’s support of abortion “rights” appealed to cultural liberals who feared that Reagan and Bush nominees would overturn the Supreme Court’s Roe v. Wade decision (though they declined to do so in the Casey decision that June). Clinton’s articulateness appealed to the college-educated, while his southern accent and that of his equally young vice presidential choice, Al Gore, enabled Democrats to carry six of the fourteen southern states. This new pattern of support for Democrats would grow larger when Clinton was reelected in 1996, and in some cases even more so for other Democrats in the twenty-first century.

Within only a few months of when Democrats led by Bill Clinton ended the Republican lock on the presidency, Republicans led by Newt Gingrich ended the Democratic lock on the House of Representatives. Clinton and Gingrich were both Baby Boomers, as defined by William Strauss and Neil Howe in Generations. Both first ran for Congress unsuccessfully in 1974, and both achieved national notice in 1978 as Clinton was elected governor of Arkansas and Gingrich as the only Republican congressman from Georgia. Both were innovative thinkers and original strategists, politicians of flexibility sometimes indistinguishable from opportunism, but also capable of learning from mistakes. By the middle 1980s, Clinton’s skills and connections made him well known to political insiders, while Gingrich’s gadfly activities in the House, amplified by C‑SPAN, made him well known among congressional conservatives and correspondents.

By 1984, as Reagan was sweeping to reelection, Gingrich began predicting that Republicans would capture a majority in the House; he was wrong through several electoral cycles, but identified the reasons they eventually did. The Watergate–era Democrats, whose abilities to hold conservative districts may have saved Democrats from losing their majorities several times, would either retire or run for other offices or die. Southern Democrats would either change parties (as Phil Gramm and Kent Hance did after losing committee posts for supporting the Reagan tax and spending bills), or retire, or die, or (occasionally) be defeated, and would be replaced in almost every case by Republicans more in line with their districts on national issues. Third, when Democrats did have the presidency as well as Congress and could attempt to pass their liberal policies into law, as they were not able to do explicitly when Reagan and Bush were president, they would be unpopular enough to create Republican majorities.

Some of these things began happening in the 1980s: retirements and deaths, a few defeats, an ebbing of Democratic support in the South. In 1989, when Dick Cheney resigned from the House to become secretary of defense, Gingrich was elected to succeed him as minority whip—by just two votes. But his election undercut the minority leader, Robert Michel, and legitimized Gingrich’s role as the House Republicans’ chief electoral and policy strategist. In 1990, even as Republicans lost a little ground nationally, they won a higher percentage of the House popular vote in the South than in the North, for the first time since Reconstruction; they have continued to be stronger in the South in every election since. Finally, in 1994 the correlation of forces for Gingrich’s revolution fell into place. Hillary Clinton’s failed health–care proposals seemed to many voters a revival of big–government policies they thought they had left behind, while Clinton’s tax increases passed by Democratic votes antagonized high-income suburban voters. Gingrich recruited candidates all across the country and got them in September to endorse a Contract With America, pledging mainly procedural reforms. On Election Day, Republicans won a majority in the Senate and gained 54 seats in the House, for their first majority (230‑205) since 1952.

As Speaker-designate, Gingrich pushed through the Republican Conference a series of party reforms eerily similar to those pushed through the Democratic Caucus by Phil Burton twenty years before. Chairmen of committees and Appropriations subcommittees would be elected by a steering committee, on which the Speaker, the majority leader, and the whip would have multiple votes and choose other members. Term limits of six years would be imposed on chairmanships, assuring constant rotation in office and incentives for would-be chairmen to raise money for colleagues’ campaigns. Just as Burton’s reforms had created the reliably liberal party that Franklin Roosevelt called for, Gingrich’s reforms created the reliably conservative party that Roosevelt foresaw as its inevitable opposition.

By 1996 the patterns that would prevail for at least two decades were largely set. Clinton was reelected with 49 percent of the vote, and Republicans won the popular vote for the House of Representatives by a 49 to 48 percent margin. In three of the next five presidential elections, Democrats received between 48 and 51 percent of the votes and Republicans between 46 and 51 percent. In ten of the thirteen House elections from 1994 to 2016, Republicans received between 48 and 52 percent of the popular vote, and Democrats between 44 and 49 percent. There was a decided move away from Republicans and toward Democrats that produced a Democratic House majority in 2006 and a 53 to 46 percent victory for Barack Obama in 2008, plus what amounted to Democratic supermajorities in Congress in 2009–10. But in 2010 and 2012, the 1994–2004 patterns were back in place. This was a period of persistent partisan polarization and parity of a length unprecedented in American history, following a quarter century of ticket splitting and vast partisan swings.

In this period, cultural issues—on many of which there had never been partisan dispute, most notably abortion—were determinants of partisan allegiance, and the demographic factor most highly correlated with voting behavior was religion, or degree of religiosity, with the most religious in each sectarian group leaning Republican, and the least religious leaning Democratic. The Republican core group was no longer northern white Protestants—many of the more affluent among them now voting heavily Democratic—but could be described as white married people, with a heavier concentration in the South than in the North, and more in areas well outside the million-plus metropolises than in their affluent suburbs. The Democratic Party depended increasingly on nonwhites, particularly blacks but also Hispanics and Asians; and on white college graduates, especially single women. In the 1940s and 1950s, affluent Americans outside the South (and increasingly to some extent in the South as well) voted overwhelmingly for Republicans, and in presidential and many other elections they would continue to do so by lesser margins until the 1990s. But as cultural issues came to overshadow economics for these and other voters, affluent and college-educated Americans trended Democratic, to a point that would have astonished the New Dealers: The 2008 exit poll showed Barack Obama carrying voters in the highest ($200,000+) income bracket over John McCain, and the 2016 election returns showed Hillary Clinton carrying many affluent neighborhoods by more than two-to-one margins over Donald Trump.

Because many of these Democratic-trending groups were becoming a rising percentage of the national electorate, predictions were made of an “emerging Democratic majority”: in the 2002 book of that name by Ruy Teixeira and John Judis, in the journalism of Ronald Brownstein in the National Journal and the Atlantic, in the pollster Stanley Greenberg’s book Ascendant America. The demographic clustering of these heavily Democratic groups in central cities, sympathetic suburbs, and university towns gave Democrats an advantage in the Electoral College, but gave Republicans an advantage in congressional and legislative elections in equal-population districts.

The best guide to political predictions often seemed to be the past election: in the still–fluid politics of the early 1990s, five states switched 57 electoral votes between the 1992 and 1996 presidential elections; between 2000 and 2004, only three states with 16 electoral votes did so; between 2008 and 2012, the switchers were two states with 26. And the focus of the Electoral College shifted as well. In the first half of the twentieth century, New York and some large states that resembled it cast the key electoral votes in closely contested elections. From the 1960s through the 1980s it was the South—in particular, white southern conservatives, who swung the key electoral votes in such races. By the end of the 1990s and into the twenty-first century, the key electoral votes were cast in the industrial Midwest and the kindred state of Pennsylvania, and most visibly in Florida, where many voters had grown up in the aforementioned states, or in New York or the South when they were successively the fulcrum point of American politics.

Donald Trump’s victory in the 2016 presidential contest, though almost entirely unpredicted, did not represent a major change from the polarized partisan patterns. Trump ran significantly better than previous Republicans among white non-college–graduates and significantly worse among white college graduates. The differences were not overwhelming, however, and in terms of popular votes those differences tended basically to cancel each other out. He lost the popular vote by 2 percent, the average for Republican nominees in the previous four presidential races. But the electoral vote count was different. Trump’s weak showing among white college graduates resulted in worse Republican showings in states including California, Arizona, Colorado, Texas, and Georgia—worse percentages, but no loss of electoral votes as compared with 2008 or 2012. In effect there was a movement toward Democrats among the college-educated in metro Atlanta, Dallas, Houston, and Phoenix, similar to what occurred starting in the mid-1990s in most non–southern metro areas.

But Trump’s relatively strong showings among white non-college-graduate voters in Florida, Pennsylvania, Ohio, Michigan, Wisconsin, Iowa, and the 2nd congressional district of Maine—voters whom Democratic strategists, in their concentration on “ascendant” nonwhites and young singles, took for granted—resulted in a net gain of 100 electoral votes over 2008 and 2010 for Republicans. Hillary Clinton ran far behind Barack Obama’s showings in the nonmetropolitan parts of these states. That may owe something to the contempt she showed for “deplorables.” But the Obama benchmark may have been unusually high, particularly in large parts of the Midwest where there were traditions in both parties of support for black rights and aspirations—among Republicans, dating from the foundations of their party in the 1850s; among Democrats, dating from the rise of the industrial unions in the 1930s. In any case, a small shift in votes by members of demographic groups—similar in magnitude to the shifts among such groups between the postwar close elections of 1960, 1968, and 1976—had large consequences in personnel and policy. And if Donald Trump seemed to diverge from other past Republican nominees and presidents on issues like trade and foreign policy, such divergences have not been unknown in the history of either of America’s major parties.

Whatever the divergences of Donald Trump from other Republican nominees—and they do not seem historically unprecedented—they did not prevent him from winning near–unanimous support from Republican voters. The force of partisanship—the residual strength of the two ancient American parties—seemed undiminished and perhaps stronger than ever.

Trump’s election may have exacerbated the bitterness of political rhetoric, but it did not mark the beginning of the fulfillment of Franklin Roosevelt’s dream of two ideologically distinct political parties. The prayers of the postwar political scientists have been fully answered—to the regret of the large majority of today’s political scientists and commentators and, perhaps, practitioners. Whether Trump will be reelected is unclear as this is written; how and when post-Trump Democrats will govern remains utterly unknown.

It bears remembering, amid the harsh partisan clangor, that for most of the forty years from 1952 to 1992, when Republicans dominated presidential elections, Democrats held majorities in Congress, and that for most of the years since 1992, when Democrats have dominated presidential elections, albeit by smaller margins, Republicans have held majorities in Congress. If Americans from the Civil War up through World War II were inclined to entrust the executive and legislatives branches to one party, since World War II they have usually been inclined to entrust one branch to one party and the other to the other. But through it all the parties have been a force for stability, for channeling public opinion and enthusiasm and discontent into relatively familiar and navigable channels. Rather than regret our partisanship and disparage our two ancient political parties, we should seek to build on their strengths and cherish their formidable heritages.