I can see the headline already: "Biden Defeats Trump."
With just two days remaining before the final votes are cast, President Donald Trump's obstacles to re-election look insurmountable. The pandemic he wished would miraculously go away is entering its third wave. The economy is recovering, but after a savage recession. He is two points further behind in the polls than John McCain was in 2008 and almost as far behind as George H W Bush was in 1992. In recent columns, experienced pundits from Peggy Noonan to Andrew Sullivan to Charlie Cook — none of them card-carrying Democrats — have dared to contemplate a landslide victory for Joe Biden.
My Halloween treat for one and all is 72 years old, dating back to just before Biden's sixth birthday. It is a newspaper front page, dated Nov. 3, 1948, and it carries the immortal headline, "Dewey Defeats Truman."
Like Biden, Thomas E. Dewey was the clear Establishment pick from the start. The governor of New York since 1943, Dewey had unsuccessfully run for president twice before (failing to secure the Republican nomination in 1940, and losing to Franklin Roosevelt in 1944). He represented the moderate Eastern wing of his party, and in 1948 had fended off more ideologically right-wing primary challengers such as Senator Robert A. Taft of Ohio and General Douglas MacArthur.
Like Biden, Dewey therefore entered the race as a known moderate with ample public-policy experience. Mindful of California's electoral importance, he picked a Californian running mate, Governor Earl Warren. Like Biden, Dewey ran a low-risk, low-profile campaign, assuming the inevitability of his victory, on the advice of his highly experienced campaign managers.
As today, the mainstream media overwhelmingly favored the challenger. Dewey gained endorsements from almost every major newspaper in America. In its final pre-election edition, Life magazine featured a glossy photo of Dewey and his staff with the caption, "Our Next President Rides by Ferryboat over San Francisco Bay."
Newsweek surveyed 50 political experts on Oct. 11 and received unanimous confirmation that Dewey would win: "The landslide for Dewey will sweep the country." Drew Pearson of the Washington Post "surveyed the closely-knit group around Tom Dewey who will take over the White House 86 days from now."
As today, too, the press loathed the incumbent. The Detroit Free Press dismissed Harry S. Truman as "intellectually unqualified" to remain president. The Los Angeles Times called him "the most complete fumbler and blunderer this nation has seen in high office in a long time." Alastair Cooke of the Manchester Guardian published an article on the day of the election entitled "Harry S. Truman: A Study of a Failure."
NBC News constructed a cardboard model of the White House with two elephants that would emerge when Dewey's victory was called. As Truman's defeat was considered certain, nobody bothered to place any Democratic donkeys on the set. "What's the Use of Going Through with the Election?" was the caption of a typical cartoon two weeks before Election Day.
A significant reason for Dewey's complacency was that the pollsters gave Truman no chance. Going into the 1948 election cycle, the president's approval rating hovered in the low- to mid-30s (below Trump's 35% approval nadir in December 2017). In the months before the 1948 election, the polls gave the challenger comfortable margins of 5 to 15 points.
On Sept. 9, pollster Elmo Roper announced that "Thomas E. Dewey is almost as good as elected … I can think of nothing duller or more intellectually barren than acting like a sports announcer who feels he must pretend he is witnessing a neck-and-neck race." The Crossley Poll of Oct. 15 placed Dewey ahead in 27 states. Truman came second in all nine of the Gallup Poll's post-convention surveys.
What should have worried Dewey was that his lead in the Gallup poll shrank from 12 points in late August to 9 points by mid-October. Gallup's final voter survey a fortnight before the election showed him ahead by just 6 points. Yet contemporaries still assumed that was a sufficient margin to assure him of victory. Dewey had summed up the prevailing view in 1944: "Never argue with the Gallup Poll. It has never been wrong and I very much doubt that it will ever be."
According to David McCullough's biography of Truman, the betting odds against the incumbent were 15 to 1 on average — in some places, 30 to 1. Truman's wife, Bess, did not think Harry would win re-election. Truman was politically finished — hence the pun: "To err is Truman."
Like Trump, Truman had lost control of the House of Representatives two years previously. Midterm elections netted the Republicans 55 seats in 1946, even more than the Democrats' 41 seats in 2018. Like Trump, Truman had bitter opponents within his own party's elite. Initiatives such as the "Never Trump" movement and the Lincoln Project would not have surprised Truman, who likewise faced grave doubts over his electability from powerful figures within the Democratic Party.
Indeed, the intraparty schism was worse for Truman than for Trump. Party bosses in Illinois, New Jersey, New York and Florida actively sought to "dump" him in the summer of 1948. Progressive Democrats, getting cold feet over Truman's record on civil rights and the Cold War, defected with former Vice President Henry Wallace to form the Progressive Party, while Dixiecrats rallied around Senator Strom Thurmond and the States' Rights Democratic Party.
Like Trump, Truman also confronted institutional "deep state" opposition. FBI Director J. Edgar Hoover volunteered the services of the bureau to aid his opponent's campaign. Truman's dirty laundry was scant, yet the studies Hoover prepared for Dewey examined Truman's early ties to Kansas City boss Jim Pendergast for criminal wrongdoing. As one of Hoover's assistants put it: "No one in the Bureau gave Truman any chance of winning."
What was it that secured Truman's victory and consigned the Chicago Daily Tribune's headline writers to eternal obloquy?
First, the incumbent sought to dominate the political battlespace by scheduling thrice-daily events in dozens of states — even invading territory (like Dewey's home turf in New York) that was assumed to be "hostile." Airport rallies and whistle-stops in battleground states (California, Illinois, Ohio and Pennsylvania) leveraged the incumbent's "populist everyman" image.
Second, Truman's attacks on his opponent, his party and its institutional backers were direct, personal and visceral, earning him the censure of the press and "respectable" politicians alike who recoiled from his "ungentlemanly" methods. A New York Herald Tribune reporter scathingly characterized Truman's campaign as "bunkum piled higher than haystacks, and demagoguery tooting merrily down the track."
But people were listening. "You are the government," Truman told rapt audiences in cornfields and dusty rail towns. "I'm calling this trip a crusade. It's a crusade of the people against the special interests, and if you back me up we're going to win …"
Covering 21,928 miles by train and delivering 275 speeches (largely in modern-day "flyover country"), Truman appealed to the so-called forgotten man neglected by an out-of-touch Washington establishment. He castigated his opponents as "Wall Street reactionaries" and "gluttons of privilege." Finally, Truman directly attacked the pollsters. On Oct. 29, he told a roaring crowd in Queens to "throw the Galluping polls right into the ash can."
Truman's crushing 303-189 Electoral College victory rode on razor-thin margins in a handful of swing states. A shift of only 3,500 votes would have flipped Ohio to Dewey; of 9,000, California; of 16,500, Illinois. A shift of a mere 29,000 votes of 10,661,000 cast (0.27%) in three critical states would have handed Dewey electoral victory. In terms of the popular vote (49.5% to 45.1%), Truman's 11th-hour upset was the closest result since 1916.
Truman held together the New Deal coalition of urban Catholic, Jewish and organized-labor voters, and integrated the Northern black vote as a new pillar of Democratic Party support. He also fared unexpectedly well with Midwestern farmers, who had largely supported Dewey's first presidential bid in 1944.
Remarkably, the three-way split in the Democratic base did not scupper Truman. The Progressive and Dixiecrat parties sliced off 5.5% of the popular vote and 7.5% of the electoral vote — a standout electoral performance for minor parties, but ultimately insufficient to swing the election.
The GOP's 1948 election postmortem highlighted — as for the Democrats in 2016 — the perils of ebullient polling: In the words of historian Simon Topping, "Predictions of a Dewey landslide may have actually deterred many Republicans from voting." But turnout was low on the Democratic side, too. Contemporary political analyst Samuel Lubell concluded that "far from costing Dewey the election, the stay-at-homes may have saved him almost as crushing a defeat" as Republican nominee Alf Landon suffered in 1936, because higher Democratic turnout would have widened Truman's margin of victory.
Dewey's inability to convert his "unity" platform into votes on Election Day is a cautionary tale for the Biden campaign. In his book "Out of the Jaws of Victory," the journalist Jules Abels noted that Dewey's pitch for unity was "an issue which had no visceral appeal to the average American. It was hard to understand what Dewey was driving at. Sometimes it seemed that he was asking Americans to achieve unity by being united behind him."
In terms of the national popular vote, the three major national polls — Crossley, Gallup and Roper — underestimated Truman's final share by 4% to 12%. An election postmortem of America's leading pollsters concluded that the surveys "left relatively untouched important problems of campaign effort, turnout, state and local politics, and voting behavior as well as the more psychological problems of the formation of preferences, the relation of intention to actual behavior, and the more subtle processes of interviewing." The pollsters oversampled college-educated voters, who heavily favored Dewey, and undersampled working-class voters, who heavily favored Truman.
Insufficient polling at the state level also played a part. Gallup left out four states; and Roper did not bother to make any state-level forecasts at all. The average of state errors undershot Truman's election result by 4% to 5%.
Yet Truman won not just because of polling errors, but because of late-stage momentum. The effective ground game of the Truman campaign was crucial in the final days of the race. American elections before the 1960s were highly labor-intensive, with the door-knocking and get-out-the-vote efforts coming largely from volunteer labor — for Democrats, from organized labor; for Republicans, from the Chamber of Commerce. Truman's last-ditch effort to get out the vote mobilized infrequent Democratic voters and scooped up undecided voters and Republican defectors, since (as the pollsters' postmortem ruefully noted) "about 1 voter in 7 made up his mind within the last two weeks and three fourths of them voted for Truman."
All of this is not to predict that 2020 will be 1948 revisited. As well as the obvious similarities, there are many differences. To name just two, there are fewer undecideds going into the final phase of this election, and there has been unprecedented early mail-in voting.
But revisiting 1948 provides a salutary reminder that even bigger upsets have happened in the history of American presidential elections than the one that elected Donald Trump president in 2016. His re-election would be a surprise right up there with 1948 — or with 1992 in the UK, where the Conservatives under John Major defied a recession to defeat Labour for the fourth consecutive election.
The final polls that year gave Labour a 1.5% lead; Major's Tories won by 7.6% — a nine-point miss by the pollsters. (I wasn't around in 1948, but I vividly remember the stunned facial expressions of my left-leaning Oxford colleagues 28 years ago. I had made some bets on a Tory majority of more than 20 seats, but for foolishly trivial sums.)
Perhaps the probability of a Trump win in 2020 is only 11%. But 11% is not zero. The average American's probability of catching Covid-19 this year has been much lower (less than 3%). How much do you worry about the coronavirus?
Last week, in what the pundit Mark Halperin calls the "significant six" states — Florida, Georgia, Iowa, North Carolina, Ohio and Texas — Biden was ahead in just three, and his leads were within the margin of error. According to RealClear Politics, Trump support was growing in Florida, Georgia, North Carolina and Ohio. There were also signs of a late Trump rally in Pennsylvania — I expect him to win in Scranton, Biden's birthplace — and Arizona. If Trump wins all eight of those states, it doesn't matter if he loses Michigan and Wisconsin: He's a two-term president.
And if it happens, the postmortem will go like this. "Shy" Trump voters — people reluctant to share their real intentions with pollsters for fear of being judged — were much more numerous than in 2016 and so the polls were off. Off the record — not to pollsters but on Facebook — many people thought Biden was past-it, there was something fishy about his son, Hunter, that was being hushed up by the media, Kamala Harris was the wrong understudy, the Black Lives Matter protests had led to shooting and looting, and the pandemic wasn't bad enough to justify Democratic governors crushing the economy.
Hey, didn't Trump kick the virus's butt and bounce back to the campaign trail? And it was not only white working-class voters who felt this way. Appreciable numbers of Latino and black voters did, too. The gun purchase data were the only reliable poll.
Yes, as I argued here two weeks ago, I really can see the headline: "Biden Defeats Trump." Gun to my head, I'd say the probability of a 2008-style blue wave is still roughly double that of a 1948-style upset. But I can also see the despised underdog incumbent's triumphant grin as he holds that erroneous front page aloft.
Niall Ferguson is the Milbank Family Senior Fellow at the Hoover Institution at Stanford University and a Bloomberg Opinion columnist.
Disclaimer: This opinion first appeared on Bloomberg, and is published by special syndication arrangement.