Skip to main content
ABC News
Why The President’s Party Almost Always Has A Bad Midterm

One of the most ironclad rules in American politics is that the president’s party loses ground in midterm elections. Almost no president is immune. President George W. Bush’s Republicans took a “thumping” in 2006. President Barack Obama’s Democrats received a “shellacking” in 2010. President Donald Trump’s Republicans were buried under a blue wave in 2018. And the results out of Virginia and New Jersey last November suggest that a red wave might hit President Biden’s Democrats in 2022.

It’s worth digging into the data behind this rule, though, and the reasons why it so often holds true. Are Republicans really a lock to sweep the 2022 midterms? 

History certainly seems adamant that they will win the national popular vote for the U.S. House and gain seats there — although their prospects in the Senate are less certain. But as with any rule, there are exceptions. And some theories for why the “midterm curse” exists may contain some hints that Democrats may be able to hold their losses to a minimum. At this point, though, history isn’t on the Democrats’ side.


Since the end of World War II, the president’s party has consistently gotten a lower share of the national House popular vote in the midterm than in the prior presidential election. Indeed, in the 19 midterm elections between 1946 and 2018, the president’s party has improved upon its share of the House popular vote just once. And since 1994, when (we would argue) the modern political alignment took hold,1 the president’s party has lost the national House popular vote in six out of seven midterm elections — usually by similar margins (6 to 9 percentage points) to boot.

The exception in both cases was the 2002 election, when Republicans under Bush’s leadership won the popular vote by 4.6 points — 4.3 points more than they had won it in the 2000 House elections. But the circumstances of that 2002 midterm election were extraordinary: One year after the Sept. 11 terrorist attacks, Bush was still unusually popular

Overall, in the post-World War II era, the president’s party has performed an average of 7.4 points worse in the House popular vote in midterm elections than it did two years prior. Therefore, since Democrats won the House popular vote by 3.0 points in 2020, Republicans can roughly expect to win it by 4.4 points in 2022 if history is any guide.

Because of the way the House map is drawn, the House popular vote doesn’t translate perfectly to the number of seats the president’s party loses, but as a general rule, the drop in support for the president’s party does cost it seats in Congress — at least in the House. Since the end of World War II, the president’s party has lost House seats in all but two midterms: 2002 and 1998, when Republicans were seen as overreaching with their impeachment inquiry into President Bill Clinton. In the average midterm election during this time period, the president’s party has lost 26 House seats.2

If this happens to Democrats this November (who will probably go into the midterms with 222 House seats,3 just four more than a majority), they would easily lose the House. Now, Democrats may be able to minimize their 2022 losses because they won’t go into the election with a huge majority (fewer seats means fewer seats to lose), but there historically hasn’t been a strong correlation4 between the size of a party’s majority and their seat losses. Parties in the approximate situation in which Democrats now find themselves have done as well as a six-seat gain (Republicans in 2002) and as poorly as a 47-seat loss (Republicans in 1958).

Although the president’s party almost always loses seats in the House in a midterm, the pattern is a bit more inconsistent in the Senate. Since World War II, the president’s party has either gained seats on net or at least avoided losing ground in six out of 19 midterms.

This might sound counterintuitive given how often the president’s party loses ground in the House, but House elections are simply more susceptible to the national electoral environment than Senate elections. This is, in part, because all 435 seats are up in each House election, whereas only about one-third of Senate seats (and roughly two-thirds of states) are up.5 As such, the partisan makeup of those Senate seats can more strongly influence the electoral chances of the two parties. Moreover, Senate elections are statewide contests where incumbents have sometimes had a larger edge than their House counterparts, in part because a distinct personal brand can still somewhat override trends running against the incumbent’s party.

In fact, the House and Senate have moved in opposite directions three times over the past 75 years, most recently in 2018. Why the asymmetry? In many cases, the president’s party flipped seats in states the president had carried handily two years earlier. Take the 2018 midterms: Democrats had to defend 24 of the 35 seats up in 2018 (26 if you count two independents who caucused with the party6), and 10 of those were in states that Trump carried in the 2016 election. In the end, the GOP won four of those 10 seats — three of which were in deeply red states like Indiana, Missouri and North Dakota (the other was traditionally swingy Florida) — while still losing swing-state seats in Arizona and Nevada.

Looking ahead to 2022, it’s less likely we’ll see the Senate and House move in different directions, as Republicans have only two Biden-won Senate seats to defend, Pennsylvania and Wisconsin, which are states Biden won by less than 2 points, meaning Democrats have little in the way of easy pickings. By contrast, the GOP will likely have more opportunities for pickups, as they can expect to challenge Democratic-held Senate seats in battleground states, such as Arizona, Georgia and Nevada, each of which Biden won by fewer than 3 points.

If the electoral environment proves strong for Republicans, they also could capture a state like New Hampshire, which Biden won by 7 points, even though GOP Gov. Chris Sununu, Republicans’ most hoped-for Granite State candidate, announced he isn’t running. And if things go really well for Republicans in 2022, an even bluer state like Colorado, which Biden won by 13.5 points, might be in play, too. That may sound crazy, but recall that in 2010 the GOP flipped a seat in Illinois, two years after Obama had carried his home state by 25 points.


OK, we’ve told you all about how the president’s party often loses in a midterm election, but why exactly does this happen with such regularity, especially in the House? 

Political science has offered a number of explanations for what’s going on under the hood, all of which may contribute at least in part to the presidential party’s midterm curse. These can largely be grouped into three categories: a midterm “reversion to the mean” after presidential elections, a “surge and decline” in voter turnout that changes the electorate from presidential years to midterm years and a broader “presidential penalty” where the party in the White House gets punished regardless of how the country is doing.

The least convincing of these arguments is probably that midterms are simply a reversion to the mean, as it posits that the party that won the presidential election overperformed expectations and thus their performance slips as a result in subsequent elections. But if this were true, we’d expect especially strong performances by the party in the White House to be followed by sizable losses in the following midterm year — only this pattern doesn’t really show up.

While there was a 17-point swing against Democrats in 2010, after Obama won the 2008 presidential election, other presidential-to-midterm cycles directly contradict this idea. Democrats saw only a small swing against them after Clinton’s nearly 9-point victory in 1996, for instance. Similarly, Republicans experienced a 5-point backlash in 1986, after President Ronald Reagan’s 18-point landslide win in 1984.

Rather, there may be more to the idea of a “surge and decline” from presidential to midterm elections. Midterm elections consistently have lower turnout than presidential contests, and studies have found that, with all else being equal, the average voter from the out party is more likely to vote in a midterm election than the average voter from the presidential party — a concept known as “differential turnout.” 

Hatred is a powerful motivator in politics, and accordingly, those who oppose the incumbent party tend to be more motivated to show up and register their frustration with the status quo. This could be an especially strong factor when there’s a Democratic president, too, because people of color and young voters, who are disproportionately Democratic in their political leanings, are less likely to vote in non-presidential elections

We don’t know how the recent surge in Republican-backed voting restrictions might affect turnout among people of color, either. Maybe it will depress turnout because it’s harder for them to vote, or maybe it will energize turnout because they are determined not to let themselves be disenfranchised. It’s also possible Democrats’ improvement among white voters with a college degree, who usually make up a larger share of the electorate in midterms compared with presidential elections, could help Democrats at the ballot box. However, this may not do that much to mitigate Democrats’ losses if they continue to lose ground among white voters without a college degree, as seems to have happened in Virginia’s gubernatorial election last November.

But probably the most compelling explanation for the midterm curse is the “presidential penalty,” whereby some voters change their minds and vote against the president’s party. After all, midterm electorates don’t actually look that different from presidential ones, but we sometimes still see really large swings in the House vote. 

The motivation to switch sides may boil down to the concept of “balancing,” wherein a small but significant portion of the electorate chooses to vote against the White House party as a check on its power. They may perceive a Democratic president to be too liberal, for instance, or a Republican president too conservative. Regardless of the rationale, though, different studies have found evidence for balancing in both federal midterm elections as well as state-level elections.

It’s an important phenomenon, too, since voters switching which party they back can give the other party a huge boost. Consider that, if a party turns out a voter who is less likely to vote, it gains one vote on net (from 0 to +1), whereas if a party flips a voter already likely to vote, it produces a net gain of two votes (+1 for the gaining party, -1 for the losing one). Per the Democratic firm Catalist, this is what drove the blue wave in the 2018 midterms. They found that as much as 89 percent of the vote swing from Trump’s win in 2016 to the blue wave in 2018 may have come down to people switching which party they backed.

This penalty could also be broadly interpreted as a referendum on the party in power, but studies disagree as to just how much midterm results are a reaction to how people feel the country is doing. Notably, however, there is an exception to this penalty — or at least a caveat. The president’s party almost always loses seats in the House, but popular presidents have historically held their losses to a minimum. Presidents with approval ratings north of 60 percent, such as Clinton in 1998 and Bush in 2002, actually saw their parties gain House seats, while President John F. Kennedy’s Democrats lost only a handful in 1962. So there may be at least an element of referendum in the midterm result, even if it’s not the sole reason voters switch parties.

All in all, though, the takeaway from history and political science literature is clear: The president’s party is almost always cursed with midterm losses in congressional elections. This reality makes Republicans favorites to win full control of Congress in 2022 pretty much regardless of what happens over the next year — although the extent of the GOP’s advantage could grow or shrink depending on how Biden is doing as president.

Footnotes

  1. That year, more than any other, was when the South became a Republican rather than a Democratic region, kicking off the current era of close elections.

  2. Compared with how many seats the party held going into Election Day, not how many seats it won in the previous election. This is why our numbers may differ slightly from other sources. We also count vacant seats as belonging to the party that previously held them.

  3. They currently have 221 seats, and one vacant seat is extremely likely to be filled by a Democrat in an upcoming special election.

  4. R = -0.33.

  5. The 100-seat Senate is divided into three classes of seats that have staggered six-year terms. There are 33 seats each in Class 1 and Class 2, and 34 seats in Class 3, and only one of the classes is on the ballot every two years. However, in some years, there are special elections for Senate seats that aren’t normally part of a given class, so the number of Senate races can sometimes go beyond 33 or 34.

  6. Sens. Angus King of Maine and Bernie Sanders of Vermont.

Geoffrey Skelley is a senior elections analyst at FiveThirtyEight.

Nathaniel Rakich is a senior editor and senior elections analyst at FiveThirtyEight.

Comments