Monthly Archives: September 2019

HOW HOUSE ELECTIONS WILL EXPAND THE 2020 ELECTORAL COLLEGE MAP

Sign up for ESY below! We won’t bother you except to let you know about fresh, new content. We promise!


The results of the 2020 House and Senate races will largely determine the efficacy of the next president. Losing either chamber to the other party would hamstring a president’s agenda — allowing the opposition to obstruct legislation, effectively negotiate the budget, and in the case of the Senate, stall executive and judicial appointments. Given these stakes, the 2020 presidential campaigns will likely be run with an eye towards down ballot Senate and House races.

Because our political era is defined by partisanship (and straight ticket voting), what’s going on between Trump and [insert Democrat here] will probably determine who wins both the Senate and the House. The most important factor in these down ballot races will be the presidential election. In fact, no state split their Senate and presidential vote in 2016, emphasizing the importance of the top of the ticket.

Here at ESY, though, we focus on the House. So, while the Senate elections will arguably have a more clear-cut effect on presidential strategy (because measuring how a campaign targets a whole state is generally easier to see than how it targets clusters of districts), this post will focus on the House.  

According to our model, a landslide victory for either presidential candidate — say a margin of >5% in the popular vote — would all but certainly deliver their party the House. In this scenario, marginal differences in presidential strategy and which states are targeted will be unlikely to affect the topline result in the presidency or House — the party dominating the popular vote would win both. In a closer election, though, the states and voters that presidential candidates target could have down ballot affects that swing crucial House races.

The strategy to winning the presidency and House in a close election is not a straightforward one. The nuances of the electoral college and the distribution of swingy House seats will incentivize 2020 candidates to expand the presidential battleground in order to boost their party’s chances down ballot. Given that spending on the presidential general election is expected reach a record $1.7 billion, there will be plenty of campaign resources to go around. Whether a candidate like Trump would actually prioritize the Republican party over his individual electoral chances, were they at odds in any way, is unclear. Still, the fact remains that while a blinkered strategy focused on Pennsylvania, Wisconsin and Michigan might work for the Electoral College, it would hurt the party in districts crucial to both Democratic and Republican hopes of controlling the House.

There are two broad strategies that campaign can use to influence voters: resource targeting and message targeting. Campaigns can aim to influence voters with resources — primarily money and time — or with the message their campaign is sending. We’ll look at these individually and see how the campaigns could leverage them to help their friends downballot. But first, a quick overview of the 2020 Presidential battlegrounds so we can then look at how the House folds into it.

The 2020 Presidential Geographical Battleground
A quick refresher on the Electoral College in case you forgot: There are 538 electoral college votes, meaning that a candidate needs 270 to solidify a majority and the presidency. Electoral votes are distributed among the states based on each states number of members of Congress (Senators + Representatives).

According to Sabato’s Crystal Ball, both Democrats and Republicans have 248 Electoral votes in their “Safe”, “Likely”, or “Lean” categories. This leaves 42 votes in the “Toss up” category. You can see where each state falls in their map below.

I would put Florida and Michigan closer to “Toss Up” category, given that Trump won Michigan in 2016 and only won Florida by 1%, while Obama carried it in 2012 and 2008.

Regardless, given how many states are essentially off the table for either party, there are only really two parts of the country with enough flip-able states to swing the Electoral College. The first is the “Rust Belt” region in the Midwest and northern U.S.  The swingy states (Toss Up or Lean) in this category are Pennsylvania, Ohio, Michigan, Wisconsin, Iowa and — stretching the definition of the Rust Belt — Minnesota. I’ll also include Maine, New Hampshire and Nebraska’s Second District in this broad category due to many similar demographics and characteristics of these states.

Trump’s 2016 win was possible because he flipped Michigan, Pennsylvania and Wisconsin, all of which Obama carried in the previous election. These three will probably again be the most competitive states in the region, with Ohio and Iowa leaning Republican and Minnesota leaning Democratic. If 2020 repeated 2016 in the 47 states besides Michigan, Wisconsin, and Pennsylvania, these three would determine the winner. And because they have similar demographics and political environments, chances are that the candidate who wins one will win all three.

The other region that could determine the election is the “Sun Belt”, which stretches from the southeast to the southwest. The swingy states in this region are Nevada, Arizona, Texas, Georgia, Florida, North Carolina and (again stretching the definition of “Sun Belt”), Colorado. I’ll also add in Virginia, given its demographics southern characteristics.

This region, though, seems to be less of a “Toss Up” than the Rust Belt. The only state in the region that the Crystal Ball rates as a “Toss Up” is Arizona, although one could argue to put Florida in that category too. Also, North Carolina and Georgia are probably a bit swingier than Texas while Nevada is likely more competitive than Colorado.  Like the Rust Belt, these states could also determine the outcome of the Electoral College. There are many ways to split up votes, but if Democrats won Florida and Arizona (while carrying all of the states they won in 2016), they would win the Electoral College (and could afford to lose MI, WI, and PA). Other options include flipping the more challenging states of Georgia, North Carolina, and to a greater degree, Texas. There are also various scenarios in which Republicans flip Democratic states and could lose the Rust Belt.

Overall, though, the Sun Belt states are probably a bit less swingy than the Rust Belt. If Democrats were able to flip Florida and Georgia, they would probably have already carried the midwestern swing states. Similarly, if Republicans won Colorado, they most likely already have MI/PA/WI in the bag.

The specifics of the electoral calculus is less important than understanding this big point: the Electoral College hangs on these two regions of the country. And while either of these regions could hold the decisive electoral votes, due to the slightly more partisan leans of the Sun Belt states, chances are that the Rust Belt will be the decisive region.

The 2020 Presidential Demographic Battleground
The key demographic in all our Rust Belt states are non-Hispanic whites. On average, this demographic makes up 88% of these states, compared to the 69% national average. These whites are, importantly, heavily working class. Sixty-one percent of population in these states is non-college whites, relative to the 45% national average. Clearly there are differences between states — Michigan has a significant (14%) black population, largely based in Detroit, while Maine has less than 1% black voters. Overall, though, the key demographic characteristic that ties these states together is the predominance of the white working class.

The Sun Belt, on the other hand, has a large and growing minority population. Only 66% of voters in these states are white, and this number would be even lower without our expansive definition of Sun Belt that includes Virginia (69% white) and Colorado (77% white), which pull the white tally upwards. The white, non-college educated percentage is also much smaller than that of the Rust belt, at only 42% of the population. Again, there are significant differences within the region — Georgia and North Carolina’s minority population are heavily black while Texas and Florida are more Hispanic. But a growing non-white, young, college educated population is true in each of these Sun Belt states. 

The 2020 House Geographic Battleground
The districts that make up the House battleground are not as easily categorized as Electoral College swing states. Due to the fractured nature of the House, there are individual flippable districts across much the country. Taken in the aggregate, though, the primary battleground of the House is clear: the suburbs. Of the 51 House districts rated as “Toss Ups” or “Leans”, 44 of them have significant or predominant suburban characteristics according to City Lab’s Congressional Density Index. Most of these districts (40) were carried by Trump in 2016, and 21 of those Trump seats were won by Democrats in 2018. Overall, 29 are held by Democrats and 22 by Republicans.

Geographically, the seats are disproportionately concentrated in the Rust and Sun Belt. There are 17 swing seats in our (generous) definition of Rust Belt states, 12 in the Sun Belt, and 22 spread out across the rest of the nation.

The 2020 House Demographic Battleground
Unsurprisingly, the demographics of battleground House districts largely reflect the overall demographics of their states and regions. The 17 swing seats in the Rust Belt have an average of 88% white voters and 59% uneducated white voters. The Sun Belt districts are 59% white and 32% non-college whites. These numbers match up with the overall demographics of the Sun Belt and Rust Belt regions, although the proportion of non-educated whites in the Sun Belt swing districts is exceptionally low. The remaining 22 swing districts outside of these key states are in line with national averages — 74% white, 47% non-educated white.

Presidential Strategy: Resource Targeting
As discussed above, both Democrats and Republicans best chance of winning the Electoral College most likely swings with Michigan, Pennsylvania and Wisconsin. It would make sense for both presidential campaigns to pour their campaign time and resources into these three states, as they will likely determine the winner.

This strategy, however, is short sighted on two fronts. First, there is the slight chance that a Sun Belt state would flip before Michigan, Pennsylvania or Wisconsin. Second, each presidential candidate will want to boost their numbers in the Sun Belt in order to help their House teammates downballot.

The 12 swing districts in the Sun Belt will be tight and could determine control of the House, incentivizing both campaigns to invest in these states, even if their chances of flipping the states are slim. And even if the states are not likely to change hands, any presidential candidate’s improvement at the top of the ticket will help downballot, meaning that presidential campaigns could invest with the primary goal of trying to win House districts. And it is hard to imagine a campaign deciding not to invest in a state once it sees its opponent investing. If one campaign decides to fight for the Sun Belt states, we’ll probably see the other join as well.

And while this would seem unlikely in a race where candidates are strapped for cash, it’s important to remember that both candidates will have ample financial resources. This will allow the campaigns to compete in lower priority states like Texas, Georgia, and Virginia, with the dual purpose of trying to win longshot electoral college votes and helping House candidates downballot.

Presidential Strategy: Message Targeting
The other strategy, in addition to physical resource investment, campaigns will use to win over and turn out voters will be their message to voters. The tone and policies a campaign presents to voters could vary widely depending on exactly which voters campaigns are targeting.

As we discovered above, the demographics of the Rust Belt and Sun Belt are extremely different and voters in the regions will respond differently to different messages. I’m not going to prescribe messages that could work in the two regions (I’ll leave that to the campaigns), but I will describe some key demographic groups and trends that the campaigns will need to keep in mind.

  1. The Rust Belt is much whiter and less educated than the Sun Belt. We discussed this earlier, but it is a crucial dynamic who the campaigns will be targeting and the message used to target them. Is there a way for campaigns to appeal to both white working class that dominate the Midwest without turning off minorities and college educated whites in the Sun Belt?
  2. Trump’s approval with white, non-college voters is much worse in the Rust Belt, where it hovers around the low 50s, than in the Sun Belt, where it is the mid-60s. This is perhaps the most under-appreciated demographic dynamic that leaves an opening among what many consider to be Trump’s base. What message can the campaigns deliver that would win over these key voters who have soured on Trump?
  3. Minorities in the Rust Belt really hate Trump, while those in the Sun Belt are a bit more divided. This largely comes down to the president’s absolutely dismal approval (~15%) among black voters (who make up most of the minority voters in the Rust Belt and some southern states like Georgia and North Carolina) versus his marginally better approval (~25%) among Hispanic voters (who are more dominant in Florida, Texas, Arizona and Nevada). What message can Democrats use to turn out these Trump-opposed voters without firing up and helping Trump re-solidify his white non-college base (that he could be in trouble with in the Rust Belt states). And what could Republicans do to alleviate their anemic margins among non-white voters?
  4. Lastly, college-educated white voters dislike Trump everywhere, but more so in the Rust Belt where his approval is in the mid to high 30s than in the Sun Belt where it’s in the mid to high 40s. How could the parties win over these voters without jeopardizing their margins among minorities and non-college whites?

While Trump could double down on winning over non-college white voters as a strategy to lock down the Electoral College in the Rust Belt, this would pose a big problem for Republican House prospects. As discussed, a majority of House seats are in the suburbs. In the House districts located in the Rust Belt states, a Trump strategy targeting non-college whites could actually work, as they make up 59% of voters in those seats. However, this strategy could prove fatal in the Sun Belt, where the swing seats are only 32% non-college white voters. Similarly, the other 22 vulnerable districts across the nation are 47% non-college white. If Trump’s Electoral College strategy is to target his white working-class base, he will likely be surrendering many of these House seats that have majority coalitions of non-white and college educated voters.

The electoral incentives, then are for the presidential campaigns to find a message that will help them win over the crucial non-college educated white voters in the Rust Belt, but also boost their prospects in these suburban districts.  Perhaps the Democrats will employ an economic-focused campaign in the Midwest while focusing on social issues in the Sun Belt suburbs, as Obama did in 2012. Maybe Trump will tamp down on divisive immigration issues, focusing instead on economic populism, in order to help solidify his base without repelling suburban communities quite as strongly.

Wrapping Up
Regardless of the tactics and messages that the campaigns settle on, it is likely that their effect on the House elections downballot will weigh on their decision. While Michigan, Pennsylvania, and Wisconsin will probably deliver the winning Electoral College votes, the House will be decided in districts across the nation, although heavily concentrated in the Sun and Rust Belt. This complicated picture of the House will incentivize presidential candidates to compete in states and for voters that, if the House were not up for grabs, they may have ignored.

Given the importance of Congress to a president’s ability to carry out his or her vision, the presidential campaigns would be ignoring their impact on downballot House races at their peril. From what we have seen thus far, though, Trump is not the most prudent political operative. So, while the House elections may incentivize expanding the electoral map, we will have to wait and see if Trump — often obstinate and short sighted —- and the eventual Democratic nominee will take these House incentives into account or if they will focus solely on winning the Electoral College.

FORECASTING THE HOUSE: PROJECTIONS BASED ON THE PRESIDENTIAL POPULAR VOTE

Sign up for ESY below! We won’t bother you except to let you know about fresh, new content. We promise!


This is Part 3 in a short series projecting which party will win the House in 2020 given different possible presidential outcomes. Because it is a presidential election year and given the intensity of partisanship and increasing number of straight-ticket voters, the presidential results will be the best measure of the national environment and the most important factor in the House elections. This series will explore what those presidential results will mean for the House and which party can expect to win a majority of seats based on the various possibilities. You can find Part 1 here and Part 2 here.

This post is going to complete the series forecasting who will control of the House in 2020 based on the national environment, which is best measured by the presidential popular vote. Democrats currently hold 235 seats, Republicans hold 199, and Justin Amash is the lonely independent in MI-03. This means that Democrats can afford to lose, at most, net 17 seats.  This graphic, from our previous post, shows where we left off and what we still need to finish. 

The next chart shows, according to our model, the order that Democratic held seats would flip to Republicans. The model’s calculation uses the 2018 midterm margin, the changes in incumbency status from 2018 to 2020, and an adjustment for the 7.3% Democratic environment of the 2018 midterms. The last column shows the projected margin for the district in a neutral national environment, with positive numbers indicating a Republican victor and negative numbers indicating a Democratic win. Finally, the districts highlighted in red are ones where Donald Trump won >51% of the vote, which will be important for our projections below.

Safely Democratic
Let’s start with what national environment Democrats need to “Safely” hold the house. As with “Safely Republican”, I am going to build in a lot of safeguards here, setting a high bar for the House to be considered safe. These include:

  • Building in a 10 seat Democratic cushion, meaning that, for this calculation, Democrats can only lose seven seats, putting the House at 228D to 206R.
  • Assuming that Democrats lose the eight districts that Trump won by >51%.1In the “Safely Republican” calculation we assumed that Republicans lost any seat that Clinton won by >50%, not >51% like we assume here. The bar is a bit higher here because 1) There are a lot more seats held by Democrats that Trump won and Democrats are very unlikely to lose all of them. 2) A lot of these seats are in suburban zones that are swinging heavily away from Republicans.

Taking these two precautions, we already have Democrats losing eight seats — one more than our Safe model will allow. This means that they will have to win back one seat. The presidential popular vote they need to be able to expect to win this seat will be the bar they need to clear to Safely control the House.

So which House seat currently occupied by Republicans will be easiest for Democrats to flip? According to our model, the easiest seat for Democrats to flip will be TX-23.

This seat spans a majority of Texas’s western border with Mexico and is 62% Hispanic, a key demographic for Democrats. The Republican incumbent, Will Hurd, has been a uniquely strong incumbent, but barely squeaked out a 0.4% win and is not running for reelection next year. According to our calculation2Start with Republicans’ 0.4% 2018 margin. Adjust in Republicans favor 7.3% for the Democratic leaning environment in the 2018 midterms. Give Democrats a 2.7% boost due to the Republicans losing their incumbency advantage., Republicans could expect to win this seat by 5% in a neutral environment, meaning that Democrats need a 5.1% favorable environment to win this seat and be considered “Safe” to keep control of the House.

Likely Democratic
Moving to the “Likely Democratic” category, we are going to ease up on the safeguards a bit. The precautions now include:

  • Building in a 5 seat Democratic cushion, meaning that, for this calculation, Democrats can only lose 12 seats, putting the House at 223D to 211R.
  • Assume Democrats lose the eight districts where Trump won > 51%.

With these guardrails in place, Democrats can still afford to lose four more seats. The easiest of which (according to our model) would be NM-02, UT-04, CA-39, CA-21. The next seat in line, GA-06, is the one Democrats would need to hold onto. It would be 0.9% R in a neutral environment, meaning Democrats need to have a 1% advantage to make control of the House “Likely Democratic”

Toss Up
For our “Toss-Up” category we are going to find the tipping point seat — the one that would tip control of the House away from Democrats. Given the current breakdown of the House with one independent, there are actually two tipping point seats — one that would tip the House to a 217 – 217 tie and one that would tip it to a 218 – 216 Republican majority. The 18th and 19th seats that Republicans flip are these two “tipping point” seats. The 18th most likely seat to flip is Florida 27th and the 19th most likely is MI-08. These two districts would lean 1.9% and 1.4% Democratic in a neutral national environment, making these the boundaries for the toss up category.

Wrapping Up
Now that we have all the data, we can fill in the rest of our chart.  

A neutral environment, meaning the presidential popular vote is 50/50, falls into the “Lean Democrat” category. This is mostly due to their incumbency advantage in many of the most competitive districts that could determine control of the House. Democrats also have an advantage in the 1-3% margin range. If the national environment is 1%-3.3% in Democrats favor, they are “Likely” to win the House, but if it is a similar Republican leaning environment, the House a Toss Up, Lean Republican and even Lean Democrat. If either party wins by around 5% or more, though, it looks like the House is theirs.

ESY ON SABATO’S CRYSTAL BALL

Sign up for ESY below! We won’t bother you except to let you know about fresh, new content. We promise!


This week ESY is taking a brief hiatus because a piece of mine was featured on Sabato’s Crystal Ball, the University of Virginia’s Center for Politics nonpartisan political newsletter. The piece is called “The House’s Republican Bias: Does it Exist?”. Go give it a read and check back in next Sunday for a fresh ESY post. 


FORECASTING THE HOUSE: WHAT REPUBLICANS NEED TO “LIKELY” WIN CONTROL

Sign up for ESY below! We won’t bother you except to let you know about fresh, new content. We promise!


This is Part 2 in a short series projecting which party will win the House in 2020 given different possible presidential outcomes. Because it is a presidential election year and given the intensity of partisanship and increasing number of straight-ticket voters, the presidential results will be the best measure of the national environment and the most important factor in the House elections. This series will explore what those presidential results will mean for the House and which party can expect to win a majority of seats based on the various possibilities. You can find Part 1 here.

Quick Recap
If you did not read Part 1 of the article series, go read it for context. If you did (or are too lazy to go read it), here’s a quick recap. The goal is to determine what the 2020 presidential popular vote will need to be for:

  1. Republicans to be guaranteed to re-take the House
  2. Democrats to be guaranteed to hold the House
  3. Republicans to be likely (but not guaranteed) to re-take the House
  4. Democrats to be likely (but not guaranteed) to re-take the House
  5. Democrats and Republicans to be about equally as likely to control the House.

Last week’s post focused on #1 — what the presidential popular vote margin needs to be to create an environment where Republicans safely win the House. As you can see in the chart below, 4.7% is the end of “Safely Republican” territory and 4.6% is the start of the “Likely Republican” zone. Again, go read Part 1 to see exactly how this was determined.

Calculating the Likely Republican Presidential Popular Vote
Today we are continuing with the Republican side of the arch. What margin do Republicans need in the presidential popular vote to be likely to win the House?

Here are the variables we will take into account:

  • Each district’s 2018 margin. This will be the baseline for the 2020 results.
  • The incumbency advantage. FiveThirtyEight estimates that incumbents have a 2.7% incumbency advantage. So, for candidates who were incumbents before 2018, they will get no additional boost. Candidates who won an open seat, will get this 2.7% boost. Candidates who beat an incumbent in 2018 will get double the 2.7% (5.4%), to account for the incumbency advantage of their opponent that they overcame in that election.
  • Each district’s 2016 presidential vote: We will assume that the Republican loses in any seat in which Clinton won > 50%. These are highlighted in blue.
  • The 2018 midterm’s national environment. Because the environment leaned 7.3% in Democrats’ fav or, this will be subtracted from the Democratic candidate’s margin.

The chart below shows, according to the above variables, which Democratic seats Republicans are most likely to win. For this calculation, we will assume that Republicans win districts exactly in this order. Of course, this is not precisely how things would really play out, but it should, on the whole, represent what Republicans will need to win back these districts.

The districts in which Hillary Clinton won >50% of the vote, and which we therefore assume Democratic congressional candidates will win, are highlighted in blue.  

So, how many seats do Republicans need to win to fit our “Likely Republican” category. I am going to be mildly conservative in this calculation whereas for the “Safely Republican” category, I wanted a number that all but guaranteed a Republican victory, and was therefore extremely conservative. 

Here are the safeguards I am building into the calculation as to what Republicans need in order to “likely” win back the house. You can compare this to the “Safely Republican” safeguards, which were much more conservative. 

  • 5-seat cushion. Meaning I am estimating what Republicans need to win 223 seats, not the bare-majority 218.
  • Assuming that Republicans lose two of the four seats with candidates embroiled in scandal: Steve King (IA-04), Duncan Hunter (CA-50), Chris Collins (NY-27), Ross Spano (FL-15)
  • Assuming that Democrats win every seat where Hillary Clinton won >50% of the vote in 2016.

This means that Republicans will need to flip: 19 to take the majority, + 5 seat cushion + 2 to make up for seats lost due to scandal.  So, Republicans need to flip 19+5+2 = 26 seats.

The 26th seat that Democrats would flip is NY-19, which is north of New York City and encompasses much of the Hudson Valley and the Catskill Mountains.  This seat is currently held by Antonio Delgado, who beat an incumbent by 5.2%. If we adjust for incumbency (+5.4%) and the Democratic environment of 2018 (-7.3%), we find that, in a neutral environment, Delgado could expect to win his district by 3.3%. Therefore, Republican will need a 3.4% lead in the national environment, measured by the presidential popular vote, to likely win the House.

Now we can fill in the rest of the left-half of the graph. I have made a change to the graph to include a “Lean Republican” and “Lean Democrat” rating. This is the range where a party has a slight advantage, but one that could be easily overcome by just a few close districts swinging the wrong way. If Republicans win the presidential vote by in between 4.6% and 3.4%, they are likely, but not guaranteed, to take back the House. Since “Likely Republican” ends at 3.4%, we can fill in 3.3% as the start of the next zone, “Lean Republican”. The arrows indicate where new information was added.

The next post will complete the chart. We’ll calculate presidential popular vote for the Safely Democrat, Likely Democrat, Lean Democrat, Lean Republican and Toss Up categories.

FORECASTING THE HOUSE: WHAT REPUBLICANS NEED TO SAFELY WIN CONTROL

This is Part 1 in a short series projecting which party will win the House in 2020 given different possible presidential outcomes. Because it is a presidential election year and given the intensity of partisanship and increasing number of straight-ticket voters, the presidential results will be the best measure of the national environment and the most important factor in the House elections. This series will explore what those presidential results will mean for the House and which party can expect to win a majority of seats based on the various possibilities.

Sign up for ESY below! We won’t bother you except to let you know about fresh, new content. We promise!


As we get closer to election day, presidential polling will begin to expose the national political environment and how the country feels about the two parties. But because the presidential election is so far away and the Democratic primary is up in the air, the House generic ballot — which asks poll respondents to choose between a generic Republican and generic Democrat for the House — is a better measure of the nation’s feelings on the two parties than presidential polls.

Eventually, taking both of these indicators into account will give the best picture of where House Republicans and Democrats stand. As election day gets closer, I will start relying more heavily on presidential polls and start using the House generic ballot less to measure the national environment and more to understand if voters plan to distinguish between their presidential and congressional votes.

Specifically, the goal of this series is to determine what the presidential popular vote will need to be for:

  1. Republicans to be guaranteed to re-take the House
  2. Democrats to be guaranteed to hold the House
  3. Republicans to be likely (but not guaranteed) to re-take the House
  4. Democrats to be likely (but not guaranteed) to re-take the House
  5. Democrats and Republicans to be about equally as likely to control the House.

By the end of this series, the chart below will be filled. Today, though, the task is to fill in only the bottom-left portion of the graph — the presidential popular vote margin that Republicans will need to safely win control of the House. The percentages around the semi-circle represent margins over the opposing party in the presidential popular vote. 

So, what margin do Republicans need in the presidential popular vote to safely win control of the House? Heading into the 2020 elections, Republicans will need to net at least 19 seats to win back the House. The current makeup of the House is 235 Democrats, 197 Republicans, one independent and two vacancies. One of these vacancies, NC-03, is a safe Republican seat, meaning Republicans effectively have 198 seats. The other vacancy, is NC-09, will have a special election next week due to election fraud tainting the 2018 election. Lastly, the independent is Justin Amash in MI-03, who renounced his Republican identity after clashing with Trump and the party after calling for impeachment.

I am going to be extremely conservative in this calculation, because I want a number that will ensure Republicans control of the House barring wild circumstances.

As I’m being very conservative, I’m going to build in a lot of room for error on the Republican side. The safeguards are:

  • A 10-seat cushion. Meaning I am estimating what Republicans need to win 228 seats, not the bare-majority 218.
  • Assuming that Republicans lose the four seats with candidates embroiled in scandal: Steve King (IA-04), Duncan Hunter (CA-50), Chris Collins (NY-27), Ross Spano (FL-15)
  • Assuming that Democrats win every seat where Hillary Clinton won >50% of the vote in 2016.

Totaling these, in order to reach the “safe” zone, Republicans need to win 33 additional seats: 19 to take the majority + 10 seat cushion + 4 to make up for seats lost due to scandal.

To find out which 33 seats will be easiest for Republicans to flip, we will take several things into consideration. The presidential popular vote that Republicans will need to take that 33rd seat will be our margin for them to “safely” take the House.

Here are the variables we will take into account:

  • 2018 margin. This will be the baseline for the 2020 results.
  • Incumbency advantage. FiveThirtyEight estimates that incumbents have a 2.7% incumbency advantage. So, for candidates who were incumbents before 2018, they will get no additional boost. Candidates who won an open seat, will get this 2.7% boost. Candidates who beat an incumbent in 2018 will get double the 2.7% (5.4%), to account for the incumbency advantage of their opponent that they overcame in that election.
  • 2016 Presidential Vote: Again, we will assume that the Republican loses in any seat in which Clinton won > 50%. These are highlighted in blue.
  • The 2018 midterm’s national environment. Because the environment leaned 7.3% in Democrats’ favor, this will be subtracted from the Democratic candidate’s margin.

The chart below shows, in order, which Democratic seats Republicans are most likely to win, given our conditions and assumptions. The districts highlighted in blue are ones in which Hillary Clinton won >50%, which this overly-cautious projection model assumes that Democrats will win. 

The 33rd seat that Democrats would flip is TX-32 in northeastern Dallas County. This seat is currently held by Collin Allred, who beat an incumbent by 6.5%. If we adjust for incumbency (+5.4%) and the Democratic environment of 2018 (-7.3%), we find that, in a neutral environment, Allred could expect to win his district by 4.6%. Therefore, Republican will need a 4.7% lead in the national environment, measured by the presidential popular vote, to safely win the House.

Keep in mind how many safeguards were build into this: a 10 seat margin, assuming Republicans lose four incumbent-held seats due to scandal and giving Democrats every district where Clinton won >50%. Therefore, if Republicans clear the 4.7% hurdle, they have the House all but guaranteed in the bag.

So, now we can fill in the first part of our chart! The arrow points to the newly filled in projection. We know that Republican’s will take the House (barring something crazy) if they win the presidential popular vote with anywhere between a 100% margin and a 4.7% margin.

The next few weeks will be dedicated to filling in data for the rest of the chart, which will become more useful as the presidential election nears and the national environment starts to become clear.