The 2020 presidential election could be decided by one state, most likely Wisconsin or Pennsylvania. How often has that happened? Let’s take a walk through the history of presidential elections.
In 58 presidential elections since 1788, 19 — almost a third — have been close enough that the outcome would have been different if at least one key state had gone the other way. In nine of those 19, it was only one state that was big enough to make the difference. Some of those single states were very close, but others weren’t. For example, the only state in Donald Trump’s 2016 coalition big enough to decide the election for him by itself was Texas, decided by almost a nine-point margin. At the other end, five elections have been close enough that there were more than three states in the winner’s coalition that could have changed the outcome by going the other way.
The closeness of decisive states is hardest to determine in the first nine elections, from 1788 to 1820. Fewer states chose electors by direct popular vote in the earliest elections, and reliable popular-vote totals are hard to come by.
We do know that three elections in this period were close enough to be decided by a single state. The 1796 and 1800 contests between John Adams and Thomas Jefferson were both decided by margins small enough (a one-electoral-vote win for Adams, a three-electoral-vote win for Jefferson) that several states could have swung the election. In 1812, James Madison’s 19-electoral-vote margin would have vanished had he lost either Pennsylvania or his home state of Virginia. The 1812 election was the first time that New York passed Virginia as the state with the most electoral votes (the three-fifths clause held back Virginia’s electoral weight, given its significant slave population) and spelled the beginning of the end of Virginia’s dominance of presidential elections.
American elections became much more contentious and closely contested as the one-party dominance of 1800–1824 broke up and the franchise expanded to more voters. Half of the elections between 1824 and 1892 would have had a different outcome if just one key state had gone the other way. Only one of those was close enough that more than three states could have been decisive: the 1876 election, settled by a corrupt deal to end Reconstruction in exchange for counting all the Reconstruction governments’ disputed vote tallies for the Republican, Rutherford B. Hayes (including that of the most closely divided state, South Carolina).
The most closely divided state in this period was frequently Maryland, a border state that allowed slavery but stayed in the Union in the Civil War. (Maryland also often had eccentric ways of dividing its electoral votes.) Other than in 1836, when Virginia was one of three states essential to Martin Van Buren’s victory, the one or two states big enough to decide the elections in this period were always New York or Pennsylvania. The 1884 election produced the ultimate decisive state: The largest state, New York, was also the most closely divided (the two candidates were just a tenth of a point apart), it was the only state that would have changed the outcome by going the other way, and the winner of the state and the election was New York’s sitting governor, Grover Cleveland.
The era from 1876 to 1892 had seen the decline of the Lincoln/Grant Republican coalition, but William McKinley oversaw the rise of a new Republican majority that dominated American politics from 1896 to 1928, after which it was displaced by a dominant Democratic majority. Only the Taft–Teddy Roosevelt split between Republican conservatives and progressives in 1912 gave Democrats a brief period in power under Woodrow Wilson. As a result, there was only one truly close presidential election between 1896 and 1956, that of 1916, which resulted in Wilson’s re-election over a reunited Republican party headed by New York’s Charles Evans Hughes. It was the only election out of 16 in which any one state could have changed the outcome.
Hughes carried the nation’s three largest states (New York, Pennsylvania, and Illinois), so Wilson’s victory rested among a lot of medium-sized or smaller states. The largest states Wilson won were Ohio, Texas, and Missouri. Interestingly, turnout in Illinois exceeded that in New York, making this the only election from 1832 to 1964 where New York did not cast the most popular votes. Missouri and Kentucky, both border states, were the most common closely divided states in this period.
Franklin Roosevelt’s four elections, by contrast, were so decisive that his Republican opponents didn’t win enough electoral votes between 1932 and 1944 combined to win one election.
That brings us to our own era. Since 1960, six out of 15 elections — 40 percent — have been close enough to be decided by a single state. The 2000 election is the most notorious example, with not just Florida but every state won by George W. Bush being essential to his election. Four elections — John F. Kennedy’s win in 1960, Richard Nixon’s in 1968, Jimmy Carter’s in 1976, and Trump’s in 2016 — depended on one big state. California took over after 1968 as the largest state, and was the decisive element in the election of Nixon as the first Californian president.
Of course, a decisive state is not necessarily a close one. In an even more extreme example than Trump and Texas, James Buchanan’s victory in 1856 depended on his home state of Pennsylvania . . . which he won by 18.1 points. Let’s turn to the closest states that were big enough to swing an election.
The top rows on this chart just show the decisive states by size; from 1836 on, they are ranked by the closest margins among the decisive states. Only 1876 and 2000 had multiple decisive states where the winner’s margin of victory was less than two points. It’s forgotten now, but Al Gore in 2000 also won four states by less than half a point: New Mexico, Wisconsin, Iowa, and Oregon. Gore won New Mexico by 366 votes. Two of those four went to Bush in 2004. Nether Oregon nor New Mexico has been close again since Bush left office.
New York, then the largest state in the union, was decisive in three straight elections from 1880 to 1888 in which its outcome was decided by less than two points. That explains why Democrats in that era were eager to run New Yorkers such as Cleveland and Samuel Tilden. Besides Nixon, Cleveland, Buchanan, and Madison, other candidates who would have lost their races without their home states were Jefferson, Van Buren, Hayes, John Adams, and George W. Bush (twice). Wilson won by the margin of his birth state of Virginia, but lost the state he had governed, New Jersey. Gore, Cleveland in 1888, and Winfield Scott Hancock in 1880 would all have won if they had carried their home states. Trump, of course, won without even competing in his home state.
The closest state that was decisive in the 1916 race was California, with 13 electoral votes. Hughes, who had quit the Supreme Court to run, actually went to bed on Election Night thinking he had won, but late returns from the West Coast gave the race to the incumbent.
Finally we come to the tally of which states have been the most closely divided — decisive or otherwise — the most times. Twenty-nine of the 50 states have claimed that honor at least once, and, as noted above, the roll of states that have done so the most often is dominated by border states (Maryland, Missouri, Kentucky). The other frequent appearance is California, though it will likely be a very long time before that happens again.
It’s early yet to be sure of the 2020 landscape, and anyway, that’s another day’s column. But history tells us that elections often do come down to just one state, sometimes a very close one.