The 12 Biggest Presidential Landslides in American History

What counts as a landslide

The word gets used loosely. This list does not. A landslide here means overwhelming dominance on both instruments that measure a presidential election: the Electoral College and the national popular vote. The top four entries each cleared 90 percent of the electoral vote in a genuinely contested race, something that has happened only four times since organized two-party competition began. The framework pays no attention to which party a president belonged to. Franklin Roosevelt and Richard Nixon sit side by side at the top because the certified numbers put them there, not because of anything either man stood for.

One structural note matters. Electoral vote share can be inflated by a split opposition. Woodrow Wilson won 435 of 531 electoral votes in 1912, an 82 percent share, while winning just 41.8 percent of the popular vote, because Theodore Roosevelt and William Howard Taft divided the other side (National Archives, Electoral College results, 1912). That is a fascinating election. It is not a landslide of voter consensus, and it is excluded from the ranking on those grounds.

The three 49-state class victories

Three elections form the summit. In 1936 Franklin Roosevelt won 523 of 531 electoral votes against Alf Landon, who carried only Maine and Vermont. Roosevelt's 60.8 percent of the popular vote and 24.3-point margin came with roughly 27.75 million votes to Landon's 16.68 million (Clerk of the House, election statistics, 1936). In 1972 Richard Nixon carried 49 states against George McGovern, winning 60.7 percent and a raw margin of nearly 18 million votes, then the largest ever recorded. In 1984 Ronald Reagan carried 49 states against Walter Mondale and won 525 electoral votes, the highest absolute total in history, available because the College had grown to 538 votes.

Ordering the three is a judgment the methodology makes openly. Roosevelt ranks first because his electoral share, 98.5 percent, and his popular margin, 24.3 points, both exceed the others. Nixon edges Reagan for second because his popular vote margin was five points wider, 23.2 against 18.2, even though Reagan's electoral share was fractionally higher. Reasonable people could flip ranks two and three. Nobody working from the certified returns can dislodge 1936 from first.

The record holders nobody remembers correctly

Two popular vote records live outside the top three, and both surprise readers. The largest popular vote margin in American history belongs to Warren Harding, whose 1920 return-to-normalcy campaign beat James Cox 60.3 percent to 34.1, a 26.2-point spread that no candidate has matched since (Clerk of the House, election statistics, 1920). Harding's electoral share was a comparatively modest 76 percent because the Solid South still voted Democratic almost regardless of the national tide, which is why he ranks fifth rather than first.

The highest popular vote share ever recorded in a contested race belongs to Lyndon Johnson: 61.05 percent in 1964, a hair above Roosevelt's 60.8 and Harding's 60.3 (American Presidency Project, election statistics). Johnson won 486 electoral votes, but Barry Goldwater's five Deep South states plus Arizona kept the electoral share at 90.3 percent. Those five states, carried by a Republican for the first time since Reconstruction in most cases, marked the beginning of a durable regional realignment that the certified returns of the next half century confirm.

Raw vote margins tell a third story, and the record holder there is Nixon. His 1972 margin of roughly 18.0 million votes over McGovern remains the largest absolute vote gap ever certified, ahead of Reagan's 16.9 million in 1984 and Johnson's 16.0 million in 1964 (Clerk of the House, election statistics). The largest raw vote total for any candidate, by contrast, belongs to Joe Biden's 81.28 million in 2020, an election that was nowhere near a landslide, because the electorate itself had grown to more than 158 million voters. Raw totals measure the size of the country; margins and shares measure the verdict. This ranking uses the verdict.

The early republic: sweeps without contests

The gaudiest electoral percentages in the record come from elections that were barely elections. George Washington was elected unanimously twice, 69 electors in 1789 and 132 in 1792, the only unanimous choices in American history (National Archives). James Monroe won 231 of 232 electoral votes in 1820 during the one-party Era of Good Feelings; the lone dissent came from New Hampshire elector William Plumer, who voted for John Quincy Adams. The popular legend that Plumer acted to preserve Washington's unanimity is not supported by his own papers, which show he simply preferred Adams. These entries sit at ranks eleven and twelve, below smaller contested margins, because a ruler that cannot distinguish a walkover from a war is measuring the wrong thing. Thomas Jefferson's 1804 reelection, by contrast, was a real contest against a real Federalist ticket, and his 162 to 14 result earns its place in the top half on merit.

What landslides did not guarantee

The record is blunt about what a landslide buys. Nixon resigned twenty-one months after carrying 49 states. Harding died in office in 1923 with his administration's Teapot Dome scandal about to surface. Johnson's 61 percent evaporated so completely that he declined to seek renomination four years later. Hoover went from 444 electoral votes in 1928 to 59 in 1932, the fastest collapse of a landslide winner ever certified. Only Roosevelt in 1936 and Reagan in 1984 converted their blowouts into stable second terms, and Roosevelt's was immediately scarred by the failed court-packing plan of 1937.

The other pattern worth stating plainly: no election since 1984 appears on this list. Every winner from 1988 through 2024 took less than 55 percent of the popular vote, and only one, in 1988, cleared 400 electoral votes. The landslide, as a feature of American presidential politics, has been absent for four decades. If that persistence of close division produces discomfort, the discomfort belongs to the reader, not the data.