Close Presidential Elections Are Likely Here to Stay
America is going through an unusually competitive political era.
Even the most casual observers of American politics understand that this year’s presidential election is close. The latest national polling averages show Vice President Kamala Harris leading former President Donald Trump by two to three points, roughly where she has been since early August. Polls from the swing states that will decide the election also mostly show dead heats.
Something that may not be as widely known, though, is that the polling for this election is the closest on record. Inspired by a recent segment from CNN’s Harry Enten, I looked back at the Democrat-versus-Republican trial heat polls since 1936, when Gallup began surveying presidential elections. Using Gallup’s data through 2012 and polling averages for 2016 and 2020, we see that, without exception, the gap between the frontrunner and their opponent hit at least a five-point margin in every election—and often did so multiple times throughout the campaign.
This cycle is noticeably different. Since aggregators began tracking the national presidential polls back in November 2022, just after the midterm election, neither party’s nominee (or presumptive nominee) has ever held a lead of more than five points. The largest lead by any of them was Trump’s 4.3-point edge back in January of this year when President Biden was still in the race.
The reality is American elections are becoming more competitive, and it’s not just polls showing this. The previous two presidential elections were decided by razor-thin margins. In 2016, Trump won thanks to 77,744 votes across Michigan, Pennsylvania, and Wisconsin. In 2020, Biden’s victory came on the back of 42,918 votes in Arizona, Georgia, and Wisconsin.
This new era is largely a product of the country’s growing levels of geographic self-sorting and political polarization, and there are few signs that either development will abate anytime soon. Assuming we are stuck in this paradigm for the foreseeable future, there are a couple of trends we are likely to witness.
The first is a lack of overwhelming or dominant electoral wins, which were a hallmark of 20th-century elections. Especially strong candidates amassed huge margins as the electorate broadly swung to one party or the other. Some of the more impressive examples during that time included FDR in 1936, LBJ in 1964, Richard Nixon in 1972, and Ronald Reagan in 1984.
One way to view this is through the national popular vote. Consider that of the 25 contests for president that took place between 1900 and 1996, just five saw a popular-vote margin of five points or less, while in 12 elections that margin hit double digits. By contrast, in five of the six presidential elections since 2000, the winner’s popular-vote lead was under five points—the lone exception being Obama’s 7.2-point margin in 2008—and all five were among the top third of closest races by national popular vote in U.S. history.
More importantly, of course, is performance in the Electoral College, which ultimately decides the winner of the presidency. This, too, has become much closer during the last two decades. Outside of Obama’s victories, the share of electoral votes going to the winning candidate in every election since 2000 has been in the bottom third of all presidential elections. In fact, while Obama earned a whopping 67.8 percent of the electoral votes in 2008, that was only good for the 30th-highest share all time—firmly in the middle of the pack relative to all other elections.
The point is: landslide wins like those America experienced with some regularity during the last century probably will not materialize again anytime soon. Part of the reason for this is because fewer states are now truly competitive. It wasn’t that long ago that voters were still willing to split their tickets to back candidates of both parties for different offices on the same ballot. This kept the map of presidential battlegrounds much larger and more competitive for both parties.
As recently as 2008, the Electoral College landscape was extensive. It’s what allowed Barack Obama to win Indiana in his first campaign and nearly take Missouri and Montana, too. It’s also what made it possible for Bill Clinton to capture Kentucky, Louisiana, and Tennessee and for Ronald Reagan and George H.W. Bush to win Connecticut, Maryland, and New Jersey.
But a decades-long realignment between the two parties around education level and geographic place hit full stride following Obama’s first election and has ossified in subsequent years. Suddenly, states like Kentucky and Tennessee, whose voters are more working-class, culturally conservative, and rural-dwelling, began voting less Democratic. Meanwhile, states like Colorado and Virginia—both of which voted for George W. Bush twice and are home to growing shares of bachelor degree-holders and higher levels of urbanization—moved away from Republicans and are today considered fairly reliable for Democrats.
One way to discern this shift is in the Cook Political Report’s Electoral College ratings, which stem back to 1988 and gauge how competitive every state is. Across the three election cycles that closed out the last century, an average of 28 states were considered potentially up for grabs by either party. That number fell to 20 for the trio of elections at the start of the 2000s, and in the three that have taken place since 2012, the average was just 14. This year, that figure is at its lowest point on record, with just eight states expected to be in play.
The waning competitiveness of the Electoral College is also evident in the actual results. Between 1988 and 2004, at least 20 states were decided by fewer than 10 points each cycle, including an astounding 32 states in 1992, when third-party candidate Ross Perot ate into Clinton and Bush’s state margins. However, that figure has declined since 2008, hitting a low of just 14 in 2020. The most competitive states—those decided by fewer than five points—present a similar, if slightly murkier, picture. Prior to 2008, they regularly numbered in the double digits; since then, they’ve only done so once (2016), and it’s unlikely we’ll see it happen again this year.
There is another, perhaps even more vivid way to illustrate this. Between 1988 and 1996—covering just three presidential elections—nearly half (24) of all states voted at least once for each party.1 However, across the last six elections, that number is down to 15. All eight of the states (and districts) in play in this year’s election have broken this barrier already, meaning it’s highly unlikely we will see a new state added to that list.
Clearly, we have entered a new era in American politics. A more competitive national landscape, reflecting our growing partisan divisions, means elections will increasingly be decided by fewer voters in fewer states. Time will tell how long this period lasts—and whether it is ultimately a good thing for our democracy.
I intentionally cut off this period at 1988 rather than going back further, both because Ronald Reagan won nearly every state in 1984 (which would have necessarily inflated the number of states that voted for both parties during this period given Clinton’s two wins) and also to draw a starker contrast with the period of six elections that followed.
I still have some hope that, if one of the two parties is willing to cut off its extremists, downplay cultural issues in favor of economic populism, and poach millions of voters from the opposite party, they can win on the scale of Obama in 2008. If they keep it up, they can win on that scale more than once. It's tragic that all the current incentives go against that.
It's tragic that when we seem to need federalism the most, no one is talking about it.
I think a lot of the heat and friction in our current politics comes from both parties wanting to force their policy preferences on the whole country. After The Great Sorting, people have physically relocated themselves to be around people who share their policy and cultural preferences, and federal politics seems premised on either side thinking that they should be able to upend the voluntarily chosen preferences of half their compatriots.
My radical reform idea: grant cities with over 150k people the ability to become a city-state with slightly different apportionment rules, e.g. they get one senator and one congressman, and their votes count for .5 of a vote in the electoral college. People in red state cities are just like people in blue state cities - lots of "in this house" signs, pride flags, etc., why not let their votes matter instead of diluting them in a sea of red? It would also solve the issues in states like Pennsylvania where there are radically different political cultures in the cities and rural areas where rural voters feel disenfranchised because their votes are diluted by the massive city populations.