2020 Election Projections

Trigger Warning

Hey! This is a political post, about a highly controversial election. You may find some of these findings to be offensive, disagree with your beliefs, or downright scary.

These views are my own and do not represent any organization with whom I am affiliated. These comments are simply a political scientist participating in a fun “pick ’em” and trying to warn readers of a rough several weeks ahead.


It is time for the 2020 Election Day predictions…  Feel free to play along with the 2020 Pick ’em. These selections are based on a model that has evolved continuous since 2010. I have added some outcome expectation notes below which aim to answer questions from friends – especially regarding the risk of violence in the coming weeks.

Looking back at 2016, I thought that election was the most complicated election cycle. For many analysts, it was. Most incorrectly projected the race. 2020 is by far a more complex election cycle than any other that I have attempted to analyze. In order to capture the impacts of early voting, the impact of Coronavirus, changes to voter suppression efforts across the country, and unusually high early voting, I have been forced to make a series of uncomfortable assumptions in this model.

It is conceivable that Joe Biden could win an absolutely blowout since Bush trounced Dukakis in 1988 (412-125); it is also conceivable that Donald Trump could grow is 2016 margin to a number that indicates a mandate for his agenda (322-216).

My model projects that Joseph R. Biden will earn the votes to claim an Electoral College victory, 350-188, barring interference. I also estimate Biden carrying roughly 52.7% of the popular vote, which is comparable to the percentage of votes that Obama earned in 2008 or Bush earned in 1988. Biden will be the candidate to win the most votes in United States History, and may pass 70,000,000 votes earned.

Trump will likely post roughly 46.9% share of the popular vote, the worst total for an incumbent candidate since Bush in 1992 or Carter in a two-way race in 1980.

This election map is an absolute mess – especially compared to others. Notably, states with large cities are trending blue. (Cities are not an explicit variable in my model, but their impact is readily apparent on this map.)

Below I tackle key indicators for various states.

Texas* (38, tilts Biden)

Texas’ unprecedented early electoral turnout is remarkable. Adjusting county-level voting ratios to scale up from 2016 to 2020 levels, and assuming that the new voters turning break towards pro-Biden by a 5:3 ratio then Joe Biden should barely carry Texas.

This “competitive” variable that I use to accelerate vote totals for traditionally unlikely candidates who are somewhat competitive are carrying Biden to victory. This variable stems from research on independent and third party candidates. In short, as a candidate gets closer to winning, it is easier for them to surge into the lead — a trend that served as Trump sized wrecking ball into the 2016 Blue Wall.

Texas could be host to last minute election shenanigans by Gov. Abbott (R), who has already attempted to suppress early voting and has lawsuit to throw out ballots in Biden-leaning Harris County (Houston).

Florida (29 tilts Trump)

Late polls for Florida are all over the map — a similar trend to 2016. Early voting totals by party are surprisingly even, unlike several other states.

The key variable driving results in Florida is “propaganda”, which indicates the propensity for voters in a state to fall for propaganda. The lack of changes in lifestyle during the Coronavirus pandemic reflect a likelihood to accept political misinformation. Considering several reports indicating foreign influence in this election cycle for Trump, propaganda variable nudges support towards Trump.

In the off chance that Biden does carry the state, expect Florida Gov. DeSantis (R) to intervene on Trump’s behalf.

Pennsylvania (20 tilts Biden)

Most analysts suggest that Biden’s home state is critical for the election. Biden’s home state

Biden’s fracking unforced error during the second debate appears to have not boosted early voting in key parts of the state.

There are an alarming number of unreturned from the Philadelphia area, which is disproportionately low compared to other Dem-leaning, major cities in battleground states.

Pennsylvania will likely experience a “red mirage” due to their vote counting procedure. Survey results indicate that Trump voters generally plan to vote on election day, while Biden voters generally plan to vote early. Pennsylvania’s election processing system is disproportionately geared towards election-day-in-person balloting. Most counties in Pennsylvania do not start processing mail in ballots until after polls close. This time-consuming process will likely delay Democratic votes from joining the counts, artificially creating a Trump lead early in the counting process.

Trump’s team appear to have lawsuits ready to go to demand limits on mail in balloting. Expect a Trump v. Wolf (Pennsylvania’s Governor) if vote counting does not stop early.

Ohio (18 leans Trump)

Late polls have Biden narrowly ahead in Ohio. Biden’s centrist views will likely help him in Ohio, along with Gov. DeWines (R) on-again-off-again feud with Trump over Coronavirus.

Access to early voting in Ohio is a bit of a mishmash. Notably, a disproportionate share of the early balloting comes from 2016 Trump country. There is also notable under-delivery of mail in ballots in Cuyahoga (Cleveland) and Hamilton (Cincinnati) Counties.

The key Ohio variable is “trumpshame” – which is a variant on the “competitve” variable in Texas. Trumpshame comes with an added bias for states that have demographics that are traditionally Democratic (e.g., labor unions) but are competitive for Trump. Some current analysts call this a “shy Trump voter” syndrome, where voters are ashamed to say that they support. Ohio was particularly elastic to trumpshame in 2016, and I expect a similar outcome in 2020.

If Biden is close in Ohio, then Trump is in for a long night.

Georgia* (16 tilts Biden)

With two competitive Senate races, media markets that bleed into other states with key states, and notable shift towards the Democratic party in recent years, Georgia will likely log the highest ratio for dollars spent per vote cast.

Although late polls show the race as a slight lean towards Trump, I expect a “competitive” variable multiplier effect to impact the race, to the extend that Biden may win.

The 2018 gubernatorial race indicated just how vulnerable the Republican hold on the state was – with the Republican winner arguably abusing his office to unjustly impact the election result.

If Biden wins, expect Gov. Kemp (R) to cooperate with demands for recounts or disqualification of ballots from Democratic-leaning districts; also expect a brutal series of legal proceedings that drudge up wounds inflicted on the Civil Rights Act.

Michigan (16 tilts Biden)

In 2016, I used Trump’s use of “China” as a protectionist dog whistle for Michigan’s Democratic union base to defect. It appears that strategy will work. Today, Trump’s use of China reflects a xenophobic dog whistle which may still work on some Michiganders, but it will likely not carry as much weight as it did in 2016.

Considering Trump’s implicit support for a white nationalists storming the capitol building with firearms and a later planned kidnapping coup, tensions are high in Michigan.

Biden has polls at roughly 50% for months. Unless there is a sharp polling error that has been sustained for months, Biden should carry Michigan.

North Carolina (15 tilts Biden)

In 2016, I missed on North Carolina. I did not appropriately apply a voter repression variable, as I had missed just how heavy-handed North Carolinian Republicans were with their limiting African Americans easy access to the voting booth.

After a series of court cases about North Carolina’s gerrymandered Congressional map and various court cases striking down voter suppression rules, I am confident that my more muted variable to closer to accurate.

The “enthusiasm” variables are driving Biden’s projected win in North Carolina. Early voting totals skew heavily towards the Democratic Party. North Carolina’s mail in Republican totals are disproportionately low, which matches a less enthusiastic voting base found in surveys.

Arizona (11 tilts Biden)

In 2016, I missed on Arizona. I accurately projected that Clinton would overperform against polls at the time — I simply overestimated that error.

This year, Biden is ahead in the polls, and Arizona voters are breaking towards Biden on their two of their most important issues: the economy and the pandemic.

Arizona is potentially going to create a blue mirage, as many election boards are already counting early ballots. As a result, the vast majority of Arizona’s vote count could be available within minutes after the polls close at around 9p ET/6p PT. Republican-leaning in-person votes will likely close that gap as the night wears on.

Wisconsin (10 tilts Trump)

I missed in Wisconsin in 2016. I misread frustration with former Gov. Walker (R) and the resurgence of labor union authority.

Wisconsin is the reason I created the “propaganda” variable. Many low-sophistication voters exist in the state. (See Making a Murderer or the Kenosha experience the last several weeks.) These voters appear easy manipulated by false propaganda, as it is surprisingly difficult to find a mask being worn properly outside of Madison or Milwaukee.

Early voting from these areas are disproportionately from Trump country counties, which indicates a repeat surprise performance for Wisconsin.

In 2016, I suggested that “Assuming that the GOP remains in the Trumpkin stance, this map become more familiar moving forward.” I believe that the Republican dedication to Trump’s policies — specifically the kakistocratic handling of the Coronavirus pandemic — will keep parts of the Midwestern “blue wall” in play for the GOP. That said, new urban centers in the sun belt are maturing into more cosmopolitan societies with an established liberal base. These areas will likely shift the state governments towards the Democratic Party for the coming years.

Chances for Skullduggery

There is a significant chance for election irregularities during this election cycle – either during or after election day. Lawsuits to determine rules changes and rule enforcement are all but guaranteed.

Trump will very likely claim victory on Tuesday night before election boards can count and report a significant share of early ballots. Because in-person election day votes are counted at the precinct level, those results should be readily available within a few hours of the polls closing. Some states are projecting multiple days of vote processing and counting.

We are likely headed toward political violence. (More below.)

We may be headed to a Constitutional Crisis. The House of Representatives may decide to count votes that are not in line with ruling from the Supreme Court of the United States — especially if large vote margins point towards a Biden win and a party-line SCOTUS ruling throws Electoral College votes towards Trump.


Each of the 35 senate races are worth a point; the House of Representatives serves as a tiebreaker.

  • AL – Tuberville (formerly Democratic, Jones)
  • AK – Sullivan (R)
  • AZ – Kelly (D) (formerly Republican, McSally)
  • AR – Cotton (R)
  • CO – Hickenlooper (D) (formerly Republican, Gardner)
  • DE – Coons (D)
  • GA – Ossoff (D) (formerly Republican, Perdue)
  • GA – Loeffler (R), after runoff
  • ID – Risch (R)
  • IL – Durbin (D)
  • IA – Ernst (R)
  • KS – Marshall (R)
  • KY – McConnell (R)
  • LA – Cassidy (R), after runoff
  • ME – Gideon (D) (formerly Republican, Collins)
  • MA – Markey (D)
  • MI – Peters (D)
  • MN – Smith (D)
  • MS – Hyde-Smith (R)
  • MT – Bullock (D) (formerly Republican, Daines)
  • NE – Sasse (R)
  • NH – Shaheen (D)
  • NJ – Booker (D)
  • NM – Udall (D)
  • NC – Cunningham (D) (formerly Republican, Tillis)
  • OK – Inhofe (R)
  • OR – Merkley (D)
  • RI – Reed (D)
  • SC – Graham (R)
  • SD – Rounds (R)
  • TN – Hagerty (R)
  • TX – Cornyn (R)
  • VA – Warner (R)
  • WV – Moor Capito (R)
  • WY – Enzi (R)

50 Democrats +5 (45 +6 from Republicans -1 to Republicans)
48 Republicans -5 (53 -6 to Democrats +1 from Democrats)
02 Independents UC

There is plenty of noise in the Senate models as well. It is unclear of coattail effects will matter in such tight races. Using my traditional model and applying it to Biden’s expected map, we end up with a majority of Senators in the Democratic Caucus — including the two independents and a potential Vice President Harris.

House of Representatives (Tie-Breaker)

I expect modest gains for the Democratic Party in the House of Representatives.

240D – 195R


2020 is a redistricting year. Every decade, states are forced to redraw district lines based on apportionment of House seats and changes to population within the state as determined by the United States Senate.

The TEA Party 2010 election led to remarkable amounts of gerrymandering. – particularly in Wisconsin, Ohio, and North Carolina, among others. If Democrats gain control of these states, then expect a series of competitive races for the more radical right candidates in newly formed districts.

Here is a risk analysis for state legislative control changes:

  • Arizona (moderate)
  • Minnesota (moderate)
  • Iowa (low)
  • Michigan (low)
  • Pennsylvania (low)
  • Texas (marginal)

Expected Aftermath

Several friends have expressed concern about a potential civil war erupting after this election. I retrofitted an old civil war model I applied to the Middle East, which in part projected the Arab Spring.

The Civil War model produces a slight chance of civil war in the United States. Institution erosion, economic uncertainty, increased economic inequality, and increased racial tensions are the key factors impacting this model. I am in the process of introducing “kakistocracy” and “kleptocracy” values to previous civil war data; early results are promising. If I apply those half-baked numbers to the contemporary US, then the model produces a low chance of civil war.

I should be clear, I am using the Fearon and Laitin definition that a civil war is an (1) intrastrate conflict (2) among actors who seek to control the government, separate a region, or use violence to change government policies (3) that kills a total of at least 1,000 people with (4) an average of at least 100 deaths per year.

I also retrofitted an old domestic terrorism model. I had generally applied it to other countries. The results when applied to the US at the state and metropolitan levels were shocking.

My domestic terrorism model produces an extremely high chance of domestic terrorism in the United States. Recent comments by the Trump administration and various state level Republican party officials exacerbate this finding. Institutional erosion, economic uncertainty, income inequality, access to weaponry, and access to communication tools, all contribute to this finding. When I added the access to firearms to the model, it registered an almost guarantee.

I should be clear, domestic terrorism includes the use of force against civilians to intimidate or coerce a government or civilian population to achieve political or social goals. These acts can range from mass shootings (violent crime incidents where at least four people are injured or killed) to arson to organized looting to bombings.

These forecasts are exogenous to whoever wins the election. However, Trump’s erosion of institutions from 2017-present and economic policies appears to be accelerating some of these risk factors.

2 thoughts on “2020 Election Projections

Leave a Reply

Your email address will not be published. Required fields are marked *