Vote Share by State

To begin with, I will examine how two-party vote share at the state level has changed in U.S. House elections from 1948 to 2020. The following interactive map plots the Republican share of the two-party vote at the state level in U.S. House elections.

The following map does the same for Democratic two-party vote share.

There are a few important observations to make based on these maps. First, note that the Republican share of the vote is the exact inverse of the Democratic share of the vote. This is because we are dealing with the two-party vote share: Republican vote share is thus equivalent to one minus the Democratic vote share, and vice-versa.

From these maps, we can also glean broad trends in the geographic coalitions of the Democratic and Republican parties. Back in the 1950s and 1960s, the maps show how the Democrats dominated the two-party vote share in Southern states like Alabama, Mississippi, and Louisiana. Today, however, it is the Republican Party that tends to succeed in these states. The maps also demonstrate how Republicans used to be competitive in more urbanized states like New York and California, which today are Democratic strongholds.

This reversal in voting patterns has been well documented by scholars like Alan Abramowitz and Eric Schickler, who chart how the political parties underwent ideological, racial, and geographic transformation in the aftermath of the New Deal coalition. These maps thus help elucidate the historical context for the political geography of the 2022 midterms.

It is important to recognize a potential pitfall of using the House popular vote to gauge state-level partisanship. In some congressional districts, incumbent politicians will run unopposed because the district is so strongly Republican or Democratic. For example, as the maps show, Republicans won 100% of the two-party vote in South Dakota in 2020. This is because incumbent Republican Congressman Dusty Johnson ran without facing a Democratic opponent in South Dakota’s lone congressional district. Of course, in reality, not every South Dakotan voter is a Republican — Democrat Joe Biden won around 35.6% of the total vote in 2020. As we continue to explore the House popular vote, it is therefore crucial to consider how having incumbents run unopposed may skew our results and analysis.

Vote Swing by State

Calculating the two-party vote share is useful for understanding broad geographic trends. But to gain a better sense of how each party fared in a given election cycle relative to past performance, I will turn to swing. The following map plots vote swing at the state level. If a state has a positive (blue) swing value, it voted more Democratic than it did in the previous election cycle two years prior. If a state has a negative (red) swing value, it voted more Republican than it did in the previous election cycle two years prior. For instance, in 2018, Democrats won around 36% of the two-party House vote in Arkansas. But in 2020, Democrats only won around 28.5% of the vote, meaning that Arkansas swung to the Republicans by around 7.5 percentage points.

The swing map illustrates how individual parties fared in specific election cycles. In recent midterm election cycles with an incumbent Republican president, such as 2018 and 2006, the Democrats made vote-share gains in nearly every state. Meanwhile, in recent midterm election cycles with an incumbent Democratic president, such as 2010 and 2014, the Republicans made vote-share gains in nearly every state. In presidential election years without an incumbent president, such as 2008 and 2016, the results are mixed — with some states swinging Republican and others swinging Democratic.

This potentially lends credence to the conventional wisdom, expressed in FiveThirtyEight, Gallup, and other sources, that the incumbent president’s party tends to lose vote share and seats in midterm election years. Turning to the 2022 midterms, if the trend from 2006, 2010, 2014, and 2018 holds, we’d expect President Joe Biden’s Democratic Party to lose vote share across the country.

Once again, it is important to recognize the limits of using the House popular vote. In 2020, the House two-party vote in South Dakota swung a whopping 37.4 percentage points towards the Republicans. But this is simply because the Republican incumbent ran unopposed in 2020, not because South Dakota actually became 37.4 percentage points more Republican in two years.

Comparing Vote Share and Seat Share

Thus far, I’ve only considered the House popular vote. Of course, control in the House comes down to which party wins the most seats, not which party wins the most overall votes. The following map plots the share of seats won by Republicans at the state level. Just as the calculation for two-party vote share considers only votes for the two major parties, the calculation used here for seat share considers only seats won by the two major parties.

The following map does the same for Democratic seat share.

As with the two-party vote share maps, these maps of state-level seat share help illuminate the development of the Republican and Democratic geographic coalitions. Back in the 1950s and 1960s, Democrats would win most if not all of the House seats in Southern states; today, that pattern has reversed. Moreover, Republicans used to pick up a significant number of seats in the Northeastern and West Coast states; today, Democrats almost completely shut them out in these states.

Next, I will compare the state-level vote share and seat share. Vote and seat share often diverge. In some cases, this is due to intentional gerrymandering, whereby partisans draw district lines to benefit their party. But the way in which voters are naturally distributed across district lines can also serve to untether the popular vote from seat share. As Jonathan Rodden demonstrates in his book Why Cities Lose, Democratic voters are so densely packed in city centers that they tend to perform poorly in single-member plurality districts, even absent intentional gerrymandering.

The following map directly compares vote share and seat share. It displays the difference between Democratic seat share and Democratic vote share in each state. If a state has a positive (blue) value, Democrats won a greater proportion of the state’s seat share than its two-party vote share. Democrats thus benefited from gerrymandering and/or electoral geography in this state during this election cycle. If a state has a negative (red) value, Democrats won a lesser proportion of the state’s seat share than its two-party vote share. Republicans thus benefited from gerrymandering and/or electoral geography in this state during this election cycle. For example, in 2020, Democrats won around 28.5% of the two-party House vote in Arkansas. But Democrats won 0 out of the state’s 4 House districts. Electoral geography thus benefited Republicans in the state by 28.5 percentage points.

One immediate observation we can make when looking at recent election years is that partisan advantage seems to align with vote share. In other words, typically Republican states also seem to have district maps that benefit Republicans, while typically Democratic states seem to have district maps that benefit Democrats. This makes intuitive sense: decennial redistricting is often controlled by state legislatures. In red states, Republican politicians are thus able to make district maps even redder, and in blue states, Democratic politicians are able to make district maps even bluer.

From a historical perspective, note how the Republicans first gain an advantage in several states starting in 2010 and retain this advantage through 2020. This could in part be the result of the Republican State Leadership Committee’s REDMAP initiative, which sought to flip vulnerable state legislatures in 2010 in order to control the 2010 redistricting cycle. The initiative was highly successful, allowing Republicans to gerrymander a greater number of districts. This Republican gerrymandering strategy could perhaps help explain why Republicans fare better in the seat share than vote share in many states today.