Biden’s Lead Is Significant, But Fragile

With the 2020 Presidential election less than 3 months away, high quality polls are being published in large numbers, giving us a clear understanding of the current state of the race. To help sort through this mass of data, we’ve calculated polling averages in all 50 states, DC, and the nation as a whole. These averages take into account the sample size and type of each poll, and each pollsters’ past performance. In states where few polls are conducted, the data used to generate the polling average is supplemented with the national polling average plus the base partisanship of the state.

West Virginia	Trump +23.0
Wyoming	Trump +20.5
Nebraska 3rd Congressional District	Trump +19.2
Oklahoma	Trump +17.8
Kentucky	Trump +15.1
North Dakota	Trump +14.7
Alabama	Trump +14.5
Utah	Trump +11.2
Idaho	Trump +11.2
Tennessee	Trump +9.0
Nebraska At Large	Trump +8.2
Indiana	Trump +8.0
Kansas	Trump +8.0
South Dakota	Trump +7.9
Montana	Trump +7.4
Mississippi	Trump +5.7
South Carolina	Trump +5.2
Louisiana	Trump +5.0
Arkansas	Trump +4.5
Nebraska 1st Congressional District	Trump +4.3
Missouri	Trump +3.7
Alaska	Trump +3.0
Texas	Trump +0.5
Iowa	Trump +0.2
Georgia	Trump +0.2
Ohio	Biden +1.2
Maine 2nd Congressional District	Biden +2.5
North Carolina	Biden +2.7
Arizona	Biden +3.9
Nebraska 2nd Congressional District	Biden +5.2
Florida	Biden +5.6
Pennsylvania	Biden +6.6
Wisconsin	Biden +7.3
USA	Biden +7.7
Minnesota	Biden +8.0
Michigan	Biden +8.2
Nevada	Biden +8.3
New Hampshire	Biden +9.7
Virginia	Biden +10.0
Maine At Large	Biden +11.4
Colorado	Biden +12.0
Oregon	Biden +12.3
New Mexico	Biden +12.6
Illinois	Biden +14.5
Delaware	Biden +14.9
Rhode Island	Biden +16.7
New Jersey	Biden +17.6
Connecticut	Biden +18.0
Washington	Biden +20.6
Maryland	Biden +21.7
New York	Biden +22.3
Vermont	Biden +22.9
Maine 1st Congressional District	Biden +23.0
Hawaii	Biden +26.3
Massachusetts	Biden +26.4
California	Biden +27.7
District of Columbia	Biden +51.9

Biden Leads Nationally And In Swing States

Clearly, these numbers paint a very optimistic picture for Biden, with a 7.7 point national lead. This is a larger lead than Clinton enjoyed at any stage of the 2016 campaign. Additionally, Trump is behind in key swing states, with Biden up 5.6 in Florida, 6.6 in Pennsylvania, 7.3 in Wisconsin, and 8.2 in Michigan. That said it is essential to remember that systematic polling errors do regularly occur. We therefore recommend you expect these averages to be up to 3 points off in either direction. Even with this degree of uncertainty, Biden looks comfortably ahead. His lead only becomes more apparent when the averages are plotted in Electoral College terms as shown in the chart below. Trump is ahead in states worth only about a quarter of the electoral votes, and behind in states worth two thirds, with the remainder being toss-ups.

The states in this chart are plotted in order of how well Trump is polling, with his biggest lead being in West Virginia, and his lead in each state diminishing clockwise around the chart. The state at the direct bottom of the circle (6 o’clock) is then the “tipping point state”. This is a term used for the state which provides the winner with their 270^th electoral vote. Clearly there is some uncertainty, but at the moment it looks likely that the tipping point state will be one of Wisconsin, Pennsylvania, or Florida. These are therefore the most important states to be watching to determine who is going to win the election. In each of these states Biden’s edge is smaller than his national lead. This difference is Trump’s Electoral College advantage, which currently looks to stand at approximately 1.5 points. That means that Biden would need to win the national popular vote by at least 1.5 points to be the favorite to win. 1.5% is smaller than Trump’s 2.5% structural advantage in 2016, reflecting Biden’s strength as compared to Clinton with non college educated whites who dominate Midwestern swing states.

This Is A Snapshot, Not A Forecast

A crucial point to note is that these are simply polling averages reflecting current public opinion. The election is still quite some time away, leaving ample opportunity for the polls to shift. If the election were tomorrow, it would most likely be a landslide victory for Biden. The possibility that public opinion could change over time is what gives Trump a real shot at winning re-election. At this stage there are still plenty of major events that could swing the election, such as Biden selecting a Vice Presidential nominee (probably Kamala Harris), party conventions, and of course several debates.

There is also an argument to be made that this year public opinion could be more volatile than usual. Sharp shifts in the economy and the Covid-19 situation have already caused polling to fluctuate, and will likely do so again before November, potentially benefiting Trump or Biden significantly. On the other hand, one could argue that in these hyper-partisan times there are few people who are seriously going to change their minds, particularly as almost everyone in the country already has a strong and deeply ingrained opinion of the President one way or the other. So far, national polling has only varied between Biden +6 and Biden +10, with almost no good polling for Trump seen to date.

Overall, it is clear that the national environment remains about as favorable to Democrats as it has for the past 3 years, with Republicans consistently underperforming by 7-9 points. The fact of the matter is that currently the Democratic base is larger than that of the GOP, and most Independents are choosing to vote blue. For Trump to have a serious shot at re-election, something substantial is going to have to change.

Confusion As Iowa Results Delayed

Do you find the Iowa caucus system confusing? Don’t worry, so does the Iowa Democratic Party. It’s the morning after and there are still zero official results. Instead, we have chaos. The Party has failed to provide a thorough or consistent explanation for the absence of results. Reports from Iowa suggest that pretty much everything that could have gone wrong, has.

The key issue seems to be technical problems with an app precincts use to report their results. The technical helpline provided for this app is overloaded, with hold times of more than an hour and some callers being hung up on as soon as they get through to a technician. The backup system for reporting results, based on phone calls, also appears to have collapsed. Additionally there are reports of widespread confusion at caucus sites regarding the process itself, including among some organisers as a result of changes in the system since 2016.

Nevertheless, both the Buttigieg and Sanders campaigns have effectively declared victory, citing anecdotal reports from their own representatives at limited numbers of precincts, and unofficial preliminary results from <2% of precincts from the AP. Meanwhile the Biden campaign is casting aspersions on the legitimacy of the process, perhaps pre-emptively making excuses for what could have been a disappointing night for Biden. The President is also jumping in, insinuating that there exists a conspiracy in Iowa to manipulate the election results. This is one of many misinformation campaigns, with another example being the widespread retweeting of a report of Biden dropping out of the 2008 race, attempting to pass it off as being about 2020.

This is a disaster for Iowa, calling into question the caucus system and Iowa’s traditional position as first in the primary calendar. If Iowa were to switch from a caucus to a primary system, it would fall victim to a New Hampshire state law requiring the state to always have the earliest primary. This year, we can expect Iowa’s influence to be significantly decreased due to the confusion and mixed messages. This is probably good news for Biden, who was comparatively weak in Iowa, and disastrous for candidates such as Buttigieg and Klobuchar who built their entire campaigns on winning Iowa. We could now see New Hampshire become the most important state in the primary, and the power of Nevada and South Carolina magnified dramatically. Sanders particularly could stand to benefit due to his strength in New Hampshire, but this is uncharted territory, so the exact impact of tonight’s debacle is unpredictable.

For now, we recommend that you don’t take any of the currently available data seriously. Entrance polls, photos of caucuses, campaigns releasing their own biased estimates, and preliminary results from a tiny proportion of precincts are all very weak data points. We are going to have to wait for the final, official results. Thankfully, there is a paper trail for the caucuses. Therefore, if the worst comes to the worst, the Party can painstakingly recount every vote and eventually arrive at accurate, reliable results. Campaigns have been told that results should be released later today, but at this stage there’s really no telling whether or not that will happen.

The Iowa Caucuses

Today is the single most important day in the Democratic Primary. Although just 41 pledged delegates are up for grabs tonight, the results and their interpretation will have knock on effects on the rest of the country.

Iowa Predictions

Current Iowa polling shows Sanders and Biden effectively tied with 23% and 22% of the vote respectively. Buttigieg and Warren have each fallen to approximately 15%, while Klobuchar has surged to 11%. However, a naive assessment of these first choice preferences is insufficient in understanding who might win, Iowa is much more complex than that. For each local caucus, candidates who receive 15% or more of first choice preferences immediately bank those votes. However, people who initially supported candidates with less than 15% of the vote will have one opportunity to realign. They could vote for candidates who have already passed the threshold, they could simply go home, or they could support other candidates below the threshold to try and get them up to the 15% mark.

The impact of this system is that second preferences become immensely important. Bear in mind that Iowa voters are generally exceptionally well informed on the candidates due to the intensity of campaigning there, so many will have a well thought out list of preferences. Generally speaking, the field could be sorted into a moderate lane dominated by Biden and Klobuchar and a progressive lane dominated by Sanders and Warren, with Buttigieg in the middle. One would expect voters’ second choices to be from the same ideological lane as their first choice. Campaigns have explicitly reached out to one another to encourage their supporters to select each other as their second choices. The voters themselves will be in the room attempting to convince those realigning to come and support their candidate, all of which makes the process very challenging to predict.

Further complicating matters is the system of delegate allocation, with just 14 delegates being awarded based on the statewide popular vote, and the remaining 27 split between Iowa’s 4 congressional districts. Not only will the allocation be related to the candidates’ popularities in a non-linear manner due to the system of realignment and the 15% threshold, but some of the delegate allocation will be very approximate. Delegates afterall, are people, no candidate can win half a delegate, so the resolution of the system is very low. For example, the 4^th congressional district has just 5 delegates, so candidates with very different vote totals could receive the same number of delegates. Furthermore, a candidate with only a few more votes than a competitor could just pip them to the post for an extra delegate, winning far more representation despite having similar vote totals. Currently, our model is predicting that on average, Sanders will win the caucus with 14 delegates. He will be followed by Biden with 12, Warren and Buttigieg with 6 each, Klobuchar with 2, and another candidate with 1, most likely Yang or Steyer. It is worth noting that these predictions come with a very high degree of uncertainty owing to the complexity of the process.

Media Narratives

This year, for the first time, Iowa will release 3 sets of results. Historically, Iowa simply released the number of delegates allocated to each candidate. However, this year, both the first preference totals and the votes after realignment will be released. It is entirely plausible that these metrics will not all be won by the same candidate, potentially resulting in 2 or even 3 candidates declaring victory tomorrow! Whether or not that happens, the media narrative regarding the results will guide the public interpretation of them. This is not trivial, it is the public interpretation of the Iowa results, especially in terms of tactical voting and changing perceptions of which candidates are electable, which makes Iowa so important. Essentially, it doesn’t help a candidate much to win 13 delegates instead of 12, what matters is whether or not CNN declares them the winner.

Typically, the winner in Iowa receives a substantial bump in the polls. This might be more complicated this year due to potential disagreement over who the “winner” is, but remains electorally crucial. The popularity boost the winner gains is usually greater if they were not expected to win, indeed it may be more useful to consider each candidate’s performance not against one another, but against the respective expectations of each candidate. Currently, the media narratives are effectively, as ever, following intelligent averages of the polls but not accounting enough for uncertainty. Hence Sanders is broadly expected to win a narrow victory over Biden, and Buttigieg has largely fallen out of media coverage. In fact, it is still perfectly likely that any of the top four candidates could win in Iowa, although Sanders does have the best chance. High expectations of Sanders could limit the boost he receives should he win, whereas the potential upside for Buttigieg is massive, something he is counting on given his weak national polling.

Finally, candidates for whom expectations are very low can gain in the polls simply by coming second or third, exceeding expectations. This is effectively what Klobuchar is banking on. With 3 different sets of results it is also plausible that multiple candidates will be clamouring to claim second or third place. As you can probably tell, this is all becoming rather confusing and far too complicated to fit into a typical TV news segment. The manner in which journalists navigate this sea of information, and the spin candidates put on the results, will have a profound impact on the remainder of the primary.

Introducing Our 2020 Democratic Primary Model

There are a lot of conflicting media narratives about the race for the Democratic nomination. Is Biden falling in the polls? Is Warren a “front-runner”? Is Hilary Clinton about to leap into the race and cruise to victory? In such a large and chaotic field engaging in equally numerous and confusing arguments, it can be difficult to see past exaggerated news articles and the biases of those writing them. One way to decode this information is by simply ignoring all the punditry and going straight to the data. That’s why we’ve created a new probabilistic model to provide an overview of how data, not people, predict the primaries are going to go.

Our model draws on polls from dozens of organisations including national and state level research. Combined with data on historic voting patterns and the political similarities of various states, we simulate the primary, including adding random variations to polling averages to model uncertainty. After thousands of simulations, the results are averaged, allowing us to analyse results in individual states and the nation overall.

The headline is that Joe Biden is clearly in the lead. Elizabeth Warren is in second and Bernie Sanders third. That said, pollsters are coming to differing conclusions on the state of the race. Some give Biden double-digit leads on his competitors, others put Warren and Sanders neck and neck or even slightly ahead of the former Vice President. These differences are mainly due to the differing sampling techniques polling organisations employ, and it’s very difficult to say which are more representative of the truth. Our model takes an average of the results, affording more weight to pollsters with strong records and large sample sizes.

You may wonder why so few candidates are winning a significant number of delegates despite the crowded field. This is down to the all important 15% threshold. Each delegate is attached to either a particular district in a state or to the state as a whole. If a candidate fails to win at least 15% of the vote in a particular district, that candidate receives none of the district’s delegates. If they win less than 15% of the overall state vote, they win none of the statewide delegates. This makes the primary very punishing for anyone but the top three, who are the only candidates consistently polling above the threshold. The chart also demonstrates Biden’s significant lead, which appears even more pronounced when we focus in on how likely each candidate is to win the most delegates.

Although he may be winning the most pledged delegates, Biden is still struggling to win an outright majority of them. So long as this is a three horse race that could be a major issue for all the candidates. If nobody controls a majority of pledged delegates by the time the national convention in Wisconsin comes around, superdelegates will get involved. If they can’t push one candidate to an overall majority, there could be an unpredictable brokered convention. This could be quite likely according to the current data, as more than 90% of the time, nobody secures a majority in the first round of voting at the convention.

The chaotic nature of this primary doesn’t stop there though. Iowa and New Hampshire are looking to be highly competitive three or even four way races. Pete Buttigieg is gaining ground in Iowa, reaching second place with 20% of the vote in his best polls. By contrast, Biden is struggling in the two earliest voting states. He could very conceivably come fourth in Iowa and then third in New Hampshire, meaning that even if he does win this primary, it will hardly be a coronation, as he is reported to have been expecting when he joined the race. Our model does simulate voters changing their minds in response to the results of earlier primaries, with our calculations based on voter behavior in previous primaries. However, this year features an exceptionally large number of candidates, and the front-runner coming fourth in Iowa would be unusual, so there is a high degree of uncertainty involved. Therefore it’s important to keep track of simulation results in early states as well as the primary overall.

Iowa and New Hampshire are not especially diverse states. Many Democrats there are the white college-educated voters who form Warren and Buttigieg’s bases, explaining their above average performance in these early states. However, the often overlooked but still crucial Nevada and South Carolina also vote before Super Tuesday, when over a third of pledged delegates are assigned, and play a key role in shaping the narrative of the race early on. These two states are a lot more racially diverse than Iowa and New Hampshire, allowing Biden to thrive. In some polls Warren and Sanders are struggling to even reach 15% in South Carolina due to their low support among people of color, a severe problem they face in much of the South.

Overall, the race continues to be a highly competitive affair. But with many candidates struggling to qualify for the December debate it may start to become a little easier to follow. There are surely plenty of twists and turns left, such as Michael Bloomberg considering entering the race. And there’s still everything to play for as the majority of key party figures, such as Obama, are yet to issue endorsements, and most voters are still considering multiple candidates.

What The 2019 Election Results Do and Don’t Tell Us

The votes are in, and the headlines are big wins for Democrats in Virginia, where they took full control of state government, and Kentucky, where they overcame partisanship to win a gubernatorial race in one of the reddest states in the country (albeit by just half a point). But Republicans did manage to hold on to full control in Mississippi, and conservatives saw some success in ballot measures. Naturally there are already a huge variety of hot takes on what this all means, let’s break the main ones down one by one.

This was a good night for Democrats

True. Across the board, Democrats significantly overperformed the partisanship of their districts and states by an average of 5 points, with that advantage rising to about 9 points in races with no Republican incumbent. These figures approximately match other indicators of the mood of the electorate, such as the congressional generic ballot showing Democrats 6-7 points ahead. Additionally, polls pitting theoretical Democratic Presidential nominees against Trump show 5-7 point leads for Warren and Sanders, with Biden holding an even greater lead. All in all Democrats can rest easy that the national environment is distinctly blue, although not quite as blue as it was in 2018, when Democrats won the House popular vote by 8 and a half points. Furthermore, Democrats and left-leaning Independents are clearly highly motivated, displaying very high turnout for an odd year election. In future analysis we should be operating with the prior assumption that something significant has to change before Republicans can dream of winning the popular vote.

The results show Trump will lose re-election

False. Elections this far away from Presidential races simply are not good indicators of voting intention so far in the future. For example, Democrats suffered heavy electoral defeats throughout Obama’s first term, yet he went on to win re-election fairly comfortably. With Trump himself on the ballot, perhaps Republican turnout will rise to match that of Democrats. Or maybe Democrats will fail to convince the nation that Trump should be impeached, yet still go ahead and impeach him. The Democratic nominee could be perceived negatively, perhaps for extreme left wing positions in Warren’s case, or old age in Biden’s. And don’t forget the Electoral College which in 2016 gave Trump an advantage of about 2 and a half points, and the political wisdom that incumbent Presidents typically win re-election. If anything, the takeaway here is that 2020 will be very competitive.

2020 turnout will be at a record high

True (probably!) We’ve now seen exceptionally high turnout in both 2018 and 2019, with many of last night’s elections showing higher turnout than the 2014 midterms. All the data points suggest a deeply politically engaged electorate who are keen to come out and vote, especially for or against the President. Healthcare, immigration and gun control also appear to motivate voters a great deal, and there will be enormous contrast between Democratic and Republican policy on all three of these issues, regardless of who Democrats nominate. 2020 turnout in excess of 70% (compared to 55% in 2016) is entirely plausible. Who this would benefit is another question. Ethnic minorities and young voters are the traditional low turnout demographics. If increased electoral engagement is driven by these groups then Democrats could be at a big advantage. That said, every demographic has room for turnout to increase substantially. This is a case where the details are everything.

Split ticket voting is alive and well

False. Sure, Kentucky voted for a Democratic governor, yet voted against Democrats down the ballot by margins of around 20 points. But this really was an exceptional case, on par with Roy Moore’s ill-fated 2017 bid for an Alabaman Senate seat. Matt Bevin was the least popular governor in the country, with a net approval rating of about -25, and a net approval of only +11 points among Republicans. He won just 52% of the vote in his primary and is infamous for putting both of his feet in his mouth at once, for example when he “guaranteed” that children in Kentucky were being sexually assaulted as the result of a teacher strike. Overall in recent elections there is very little evidence of significant ticket splitting, and in 2020 we should expect the vast majority of voters to vote the same way in all races.

Mitch McConnell’s re-election is at risk

False. We can be fairly sure that the Senate Majority Leader won’t be losing his seat in 2020. He is certainly a remarkably unpopular senator given the partisanship of Kentucky, with approval ratings comparable to those of Bevin. But the crucial difference is that McConnell is up for re-election in a presidential year. All the media attention will be on the presidential race, and Trump’s presence at the top of the ballot will help out any struggling Republicans in red states. In 2016, every state voted the same way for the Presidency and the Senate. In 2018, we saw further evidence that party identification is becoming a much more important factor than the specific person running for office, especially in federal races. We should therefore continue to assume that so long as Kentucky is safe for Trump, it’s safe for McConnell.

One last thing: Democrats now control the entire Virginian legislative and executive branches, will they seek to aggressively gerrymander Virginian seats in their favor? Before they came into power in the state, Democrats vigorously supported a bill seeking to create an independent redistricting commission. That bill has now made quite a bit of progress, but its successful passage would prevent Democrats from capitalizing on their gains in 2019 to draw maps designed to elect as many Democrats as possible. So the question is, will they hold fast to their established anti-gerrymandering position, or will 10 years of massive electoral advantages be too great a temptation to resist?

Is The Electoral College Racist?

Back in August, Rep. Alexandria Ocasio-Cortez posted an Instagram story criticizing the Electoral College, calling it “a scam”. She went on to claim that “the Electoral College has a racial injustice breakdown. Due to severe racial disparities in certain states, the Electoral College effectively weighs white voters over voters of color, as opposed to a ‘one person, one vote’ system where all our votes are counted equally.” Unsurprisingly, her comments faced severe criticism from the right, which railed particularly strongly against the accusation that the Electoral College is racist.

So is it true that this institution makes white votes more powerful than others? Well, this is something we can test. We ran fifty-thousand simulations of presidential elections featuring a generic Republican facing down a generic Democrat. From these, we determined how much of a swing state each state is, and multiplied by their number of electoral votes to create a rating of how powerful each state is in presidential elections. Dividing this rating by the number of voters in the state gives us an index of how powerful each individual vote is. But this isn’t the full picture.

The concept of “wasted votes” is the main metric being used to assess gerrymandering today, that being the practice of drawing district boundaries to the benefit of one party over the other. Two techniques are used to achieve this: “Packing” and “cracking”. Packing is where you draw a district to lean as strongly as possible in favor of one party, packing all their voters into this one district such that the party is far less competitive in every other race. Cracking is where you spread a party’s voters out evenly between many districts such that they make up 40-45% of the population in every race. Therefore even though that party’s voters may make up almost half the population, they win none of the elections.

In the below example, green voters have been packed into one district and cracked across the rest, resulting in the purple party winning 5 out of 6 elections despite having only as many voters as the green party.

The wasted votes system counts any votes cast for a losing candidate as wasted. Therefore if most elections are being won by the same party by small margins, a huge number of the other party’s votes will be judged as wasted, effectively detecting cracking. Additionally, votes cast for the winning candidate above the 50% they need to win the election are also judged as wasted. This detects packing, as if a candidate wins a huge proportion of the vote, a large number of their votes will be wasted.

Applying this to our analysis has a profound effect on the results, leaving only the votes that make an impact on the election. For example, Republican votes in Upstate New York are discounted as these Republicans are not represented at all in the Electoral College. Additionally, many Democratic votes used to build up enormous margins in New England are treated as wasted as you don’t get any more electors for winning by 30% than by 1%. This leaves us with a clear picture of which votes count, and how much they count for. By studying the race of each voter, this information is used to calculate the average power associated with a white, black, Hispanic, and Asian vote.

The results of this are striking. On average, white votes are worth 1.13 times the average vote. Black and Hispanic votes are each worth 0.81 times the average, and Asian votes just 0.75 times. This can be attributed to a few factors. Firstly, the all important Midwestern swing states (Michigan, Wisconsin, Ohio, Pennsylvania, Minnesota, Iowa) are disproportionately white. Also, many non-white voters are packed into big cities in safe Democratic states like California, such that many of their votes are wasted running up the score in landslide elections. Additionally, many black voters are “cracked” across the South, forming large minorities which can never muster up quite enough votes to win the state.

To understand better how the disparity in representation arises, we can look at which states are the main sources of each race’s voting power. White voters are fairly evenly distributed, so the breakdown for white voters shows the main sources of their voting power being the key swing states around the country, the only surprise being the increasingly competitive Texas.

By contrast, black voters are much more concentrated in the South, and so their voting power is much more dependent on Florida, perhaps the only really competitive Southern state. However black voters are also able to exert their influence in Virginia, where since the Obama era they have formed the basis of a slim but surprisingly sturdy Democratic majority. Georgia and North Carolina also score highly as a result of large ethnic minorities coming increasingly close to flipping these states in what would be a major coup for Democrats. Interestingly, large black populations in Alabama, Mississippi, South Carolina, Louisiana and Tennessee wield very little power in the Electoral College as they almost all vote Democrat, but these states never go blue.

Hispanic voting power is even more concentrated, with five states providing over half their electoral influence. Texas and Florida each provide 16% of their voting power, with most of the key swing states having only small Hispanic populations. The importance of their role in Texas cannot be overstated. Although the Texas GOP may lose Hispanic voters by large margins almost all the time, the fact that these margins are closer to 50 points than 90 means Republicans are still able to win in this racially diverse state. In other words, it doesn’t just matter whether you win or lose a demographic, the exact margin is crucial. Arizona, a new swing state, is another key source of Hispanic voting power, and is sure to be getting a lot of attention in 2020.

Finally, many Asian voters are packed into coastal cities in deep blue states, with very little presence in any swing state, resulting in their minimal electoral influence and the absence of any concerted efforts to win them over.

This analysis suggests that AOC is probably correct at least in her assertion that due to the Electoral College, white votes count for more than those of voters of color. There is also potential for this to be reversed in the future, as increasing non-white populations in Texas, Arizona, Florida, North Carolina and Georgia could make these states increasingly competitive, and perhaps even turn them reliably blue as has happened in Virginia. But for now, tens of millions of Americans are forced to live with the reality that their vote has little to no impact, and this disempowered group is disproportionately non-white.

What The Midterms Results Tell Us About Politics

The madness of the midterms is over, and the results are pretty much as expected, with Republicans building up their Senate majority but losing the House. In fact, of the races called so far, our model correctly predicted the overall results of 97.2% of them. We predicted a Senate result of 48 Democrats to 52 Republicans, very close to the probable final result of 47-53. The House model gave a similarly accurate prediction of 239-196, against the probable final result of 234-201. At first the night looked tough for Democrats, with mixed early results in Florida, a slight surprise in the defeat of Indiana Democratic Senator Joe Donnelly, and the defeat of Amy McGrath in Kentucky’s 6th congressional district. These early results seem to have played an outsized role in forming the narrative of the election, that of a good but still disappointing night for Democrats. But make no mistake, this was a blue wave. Democrats successfully competed in suburbs to win a solid majority in the House, won the popular vote by a wide margin (about 8 points), and did surprisingly well in the Senate given that this year’s map was very good for Republicans.

The next most significant takeaway is probably that incumbency matters much less than it has in the past. America voted along partisan lines to an unprecedented extent, with very little ticket-splitting. The highest profile casualties of this were Democratic Senators in the red states of Missouri, Indiana, and North Dakota. Even in Montana and West Virginia where the Democratic Senate incumbents won, they did so by surprisingly small margins, and in Florida, which isn’t even that red, Republican Rick Scott beat Democratic incumbent Bill Nelson. By the way, watch out for Rick Scott in 2024, a popular and seasoned Governor come Senator from a swing state, whose term is ending in a Presidential year, is definitely someone to be aware of when the rumours about who’s running for President start to circulate.

Next, voter engagement was off the charts, with a turnout of 49%, the highest midterm turnout since 1914! This isn’t even much lower than the 56% turnout for the Presidential race in 2016. Of course these numbers are still very low compared to most Western democracies, but if this trend holds it seems perfectly plausible that we see extraordinary turnout in 2020 when the fate of Trump’s Presidency will truly be in play, 70% turnout begins to seem perfectly possible.

Thirdly, healthcare is by far and away the most prominent issue in politics right now. Democrats built their campaign on protecting the healthcare system, and to some extent it drove their victory, as it seems to be an argument they are winning – in a recent poll, 44% of those sampled said they trusted Democrats more on healthcare, compared to just 34% for Republicans. The next most important issue was immigration, perhaps unsurprisingly given Trump making a migrant caravan in Mexico a prominent issue during the final days of the campaign. These were followed by the economy, then gun control. It’s likely that these are the issues that will shape the elections in 2020, as politicians can see the incredible extent to which they have engaged the public. An economic downturn could spell the end of the Trump Presidency, or continued growth could form the basis of his re-election campaign. And it seems beyond doubt that the Democratic nominee will make healthcare a major talking point, and bring it up as often as they can.

Thinking about 2020, 2018 taught us a lot about what the electoral map is going to look like in 2 years time. Minnesota, Michigan, Wisconsin and Pennsylvania all veered away from Republicans, while Democrats enjoyed little joy in the Sun Belt. The Sun Belt is a collection of Southern and South-Western states which lean Republican, but Democrats dream of turning blue by turning out large ethnic minorities. The Sun Belt usually means Florida, North Carolina, Arizona, and sometimes Georgia and Texas too for the more ambitious Democrats. The best path to victory for Democrats in 2020 is now clearly via the Upper Midwestern states, which gives extra weight to Joe Biden and Amy Klobuchars’ potential Presidential campaigns, as they both have proven ability to appeal to Midwestern voters. But don’t discount Beto O’Rourke. He may have lost, but in a way that’s good for him, he now has more time to build up his campaign infrastructure, visit the early primary states, and prepare for another battle. He lost by a surprisingly small margin of only 2.6 points, and managed to electrify and thrill the Democratic base in a way that no one has really done since Obama. Add that to extraordinary fundraising abilities, and you have a very very strong contender for the Presidency.

Top 6 2020 Presidential Frontrunners

Predictions as of 10/22/2018, as Republicans take a lead in Nevada and Arizona, they look almost sure to hold on to the Senate, while Democrats’ popular vote lead sees them well ahead in the House:

The 2020 Presidential campaign began on Wednesday November 9th 2016, when America woke up to discover Trump’s victory. The race may not officially begin until late 2019, but the battle has already begun, and is sure to intensify the day the midterms are done. Since 2016, the President has made it clear that he intends to run for re-election in 2020, and a rather large number of Democrats have been setting themselves up to challenge him, not to mention several #NeverTrump Republicans. But who actually has a chance of winning?

6. Vice President Joe Biden

On first glance, Joe Biden is a strong candidate for the Presidency. He has the name recognition to cut through a Democratic nomination process that could easily involve 20+ candidates. With 7 terms in the Senate and 2 terms in the Vice-Presidency, no one could ever question his experience. Combine this with his close relation to Obama, who is now seen as practically the father of modern Democratic politics, and he seems a strong contender to win the nomination. Having done so, he could run a campaign on his traditionally moderate politics. He’s always tried to position himself as a man of the people, and his trips to Wisconsin and Michigan clearly indicate some thought about a potential 2020 strategy. Add to this the fact that he was born in Pennsylvania, perhaps the most important swing state in the country at the moment, and he may seem like the perfect answer to Trump.

But Biden has run for the Presidency twice before, in 1988 and 2008, losing horribly both times. 1988 is particularly interesting, as Biden was considered a strong candidate from the very beginning, until he was destroyed by a long string of controversies: Accusations of plagiarizing speeches from the leader of the British Labour Party, as well as Robert F. Kennedy, John F. Kennedy, and Hubert Humphrey; involvement in plagiarism whilst he was at law school; lying about graduating in the top half of his class (he came 76th/85); claiming he earned 3 degrees when in fact he only got 1; and claiming that he received a full scholarship, when in fact he only got a half-scholarship. Biden is also, to be blunt, very old, currently aged 75. This means that by the end of a two term Presidency starting in 2021, he would be 86 years old. On top of the obvious mental and physical health concerns associated with this, it’s worth noting that during the primaries for the midterms, Democrats have tended to prefer younger, female candidates, and anti-establishment candidates have also been doing a little better than expected. Joe Biden is very strong on paper, and has a perfectly good shot, but it seems as though his time has probably come and gone.

5. Senator Kirsten Gillibrand

So how about a young(er), more liberal, female Democrat? Elected once to the House and thrice to the Senate, soon to become four times as she romps to victory in her New York re-election bid, Kirsten Gillibrand may well be the new face of the Democratic Party. With the weight of the mighty New York Democrats behind her, she is a fundraising titan, with $20 million raised towards defending her perfectly safe Senate seat this cycle alone. And though she may be a full blown liberal today, at the beginning of her political career she was a much more moderate Democrat, something she could plausibly call upon once she’s got the nomination in the bag and needs to appeal to the nation as a whole. On the other hand, her ties to the Clinton family and her being a female New York Democrat may well make it very easy for her to become linked in voters’ minds to Hillary Clinton, which would almost certainly not be a good thing for her campaign. Interestingly, unlike some on this list, she hasn’t yet visited Iowa or New Hampshire, the early primary states, which is considered a key step in building up for a Presidential campaign. Even unconventional candidates like Trump visit these states well before the primaries begin, so this might indicate that she isn’t yet sure about running in 2020, hence her place near the bottom of this list.

4. Senator Kamala Harris

Who’s an even more powerful and wealthy ally in a Democratic primary process than the New York Democrats? The California Democrats of course! Kamala Harris is one of a tiny group of candidates who could have a chance at outmatching Gillibrand on fundraising, and is another comparatively young female Senator. She’s even more liberal than Gillibrand, and although she is relatively new to the Senate, this could almost work in her favour, as younger Democrats seem keen to get rid of the old guard of the Party. As if this weren’t enough, she is the best candidate on this list for appealing to the Democrats’ African-American base, which is crucial during primaries. This could also be very useful during a general election. Although many Democrats believe that their path to victory in 2020 is to regain the Upper Midwest, specifically Wisconsin, Michigan, and Pennsylvania, this is hardly their only plausible route. By motivating African-American turnout, Harris could make a serious play for Florida and North Carolina, and perhaps even make Georgia and Arizona genuinely competitive. With the Trump campaign also needing to defend slim and faltering majorities in the Upper Midwest, this strategy could make for a very strong campaign indeed.

3. Senator Bernie Sanders

To all intents and purposes, Bernie Sanders is already running for President. His 2016 campaign never truly ended, and support for him is still strong within the party. He’s visited the early primary states of Iowa, New Hampshire and South Carolina, and even has a book out this year. The real question is not whether he will run, but whether he can win. We can clearly see from the 2016 primaries that he has a lot of support, he won 43% of delegates and only narrowly lost the popular vote to Hillary Clinton. That said, Clinton was not an especially popular candidate among Party members, hardly inspiring the levels of enthusiasm Obama or Sanders generated. So if Sanders couldn’t beat Clinton, could he beat any of the candidates on this list?

Furthermore, at 77 he’s even older than Biden, there has been some talk of a one term pledge, where he promises not to run for re-election, but not from Sanders himself, and it’s likely that such a pledge would harm him during the nomination process as Democrats will be keen to get a strong incumbent into office to secure a win in 2024. In a general election, he may struggle due to being perceived as a socialist, which is still a word with strong negative associations for many Americans, particularly those in high turnout demographics. Despite all this, he is almost definitely running, he has an established base of enthusiastic supporters, and extreme views are growing more popular and electable, as President Trump shows. Sanders is going to be a political heavyweight during the primaries, and his chances should not be underestimated.

2. Senator Elizabeth Warren

Elizabeth Warren is Hillary Clinton’s natural successor, and the one to watch in the Democratic primaries. She has her finger on the pulse of the party, is generally a very skilled politician, and has incredible fundraising abilities with the support of the Massachusetts Democrats, having raised $34 million for her totally uncompetitive re-election campaign this year. She was outspoken on opposing Kavanaugh, has been a fierce Trump critic since the beginning of the 2016 campaign, and was rumoured to be a possible VP pick for Clinton. She has strong support from the left and centre of the Party membership and from across the party establishment – she even received two electoral votes for the Vice Presidency in 2016 from faithless electors. She is running, and she is utterly formidable. A political juggernaut like her on the left wing of the Party may well instantly knock Gillibrand and Sanders out of the running after the first couple of primaries, and if she manages to win the Democratic nomination she would be a similarly fierce candidate against Trump. The only conceivable mark against her is her close association with the deeply unpopular Clinton, which will likely be quietly harming her campaign throughout the process.

1. President Donald Trump

This one is obvious. We know he’s running, we know he’s so popular amongst Republican members that he’s practically guaranteed the nomination, and as an incumbent in an age of two term Presidents we know he has a good shot at winning. His unpopularity is overblown, he’s managing about 42% approval ratings, only slightly worse than Obama’s were at this point in his Presidency. Although the 2018 midterms are looking to be messy for the GOP, holding level in the Senate and losing about 40 House seats, the 2010 midterms were much worse for Obama, with the Democrats losing 6 seats in the Senate and 63 in the House. This all suggests that Trump should be just fine in 2020.

On the other hand, Obama had a lot more room for error between his campaigns. In 2008, he was elected with a 7.2 point popular vote margin and 365 electoral votes, whereas Trump actually lost the popular vote by 2.1 points in 2016 and received only 304 electoral votes. Obama could comfortably afford to lose North Carolina, Indiana, and the 2nd congressional district in Nebraska in 2012 and see his popular vote margin shrink to 3.9 points, while still winning very well. Trump has no such luxuries in 2020. He needs to hold on to very narrow margins in Florida, Michigan, Wisconsin and Pennsylvania, as well as protecting unreliable Republican majorities in North Carolina, Ohio, Iowa and Arizona. He can afford to lose one or two of these, but no more. In terms of opportunities to attack, Trump has New Hampshire, Nevada and perhaps Maine at large. With only 4, 6 and 2 electoral votes respectively, none of these are very exciting for him. However, Minnesota has a full 10 electoral votes, and although it hasn’t voted for a Republican Presidential nominee since 1972, Clinton only won it by 1.5 points in 2020, so Minnesota is very much on the table and could potentially turn the election on its head. Trump is by far and away the most likely candidate to win the Presidency in 2020, despite a variety of strong potential challengers setting themselves up to face him down, and the power of incumbency should not be underestimated.