Lastly, we will take a look at the House and Senate.
In contrast to the blue wave consensus, our data show a much more sharply divided country which is reflected in our forecasts.
In the absence of a blue wave, we see Republicans losing a net 3 seats in the Senate leaving it divided 50/50, versus consensus forecast of Democrats taking control with ~ 52 seats. For the House, Republicans should gain seats and could even come close to regaining control, versus consensus which states that Democrats will gain seats and have virtually no chance of losing control.
Our general understanding is that consensus forecasts assume a blue wave is coming which will produce net gains / victories for them across the board – Presidency, House and Senate.
Our alternative data does not show a blue wave, but instead shows many tight races in the midst of a volatile environment and a very late surge for Trump. Our longer term data (perhaps surprisingly) shows the US slowly shifting to the right, in contrast to the apparent consensus view of a leftward push.
Lower Confidence Level
Alternative data analysis tends to work better when more data is available and when time series of that data is consistent (without significant holes). For national and/or high profile elections, alternative data tends to work well. However, for smaller local elections, there are often issues with inconsistent data.
In short, our confidence level declines as data becomes spotty. However, we can still make general forecasts for the Senate and House based on limited alternative data for those races, on our Trump/Biden forecast, and on our basic political outlook.
Trump/ Biden Forecast
As a backdrop, we should highlight that we have forecasted a Trump victory with 302 electoral votes, which compares to a 350 electoral vote Biden victory based on consensus forecasts. In very general terms, consensus seems to believe that the US has shifted left and that this election will unleash a ‘blue wave’. In contrast, we see the country in the process of slowly shifting right based on a variety of alternative metrics.
Consensus Blue Wave Forecasts
Let’s start with an overview of consensus forecasts for the House and Senate. Based on an average of mainstream forecasts there is a 98% probability that Democrats will keep the House and a 74% chance they will win control of the Senate. There has also been a lot of talk of a Democrat ‘Trifecta’, or Democrats winning the Presidency, Senate and House. Such an outcome is rather unsurprising given the probabilities provided by forecasters.
These forecasts are not done in a vacuum, but also envision Biden winning the election. They place the probability of that outcome around 90%.
We believe the consensus forecasts assume a general ‘blue wave’. This is not the type of election where voters will discern between various candidates up and down the ballot. Seemingly, consensus states that there will be more straight party votes and that these will be overwhelmingly pro-Democrat. This is really the only way to interpret the forecasts given they predict the incumbent will lose, the incumbent’s party will lose control of the Senate and the opposition will expand control in the House. When you further see the probabilities attached to these forecasts (extremely high), you can really only interpret this as a wave election.
So, what would happen if Biden loses? What if this blue wave or tsunami does not occur? What if Trump wins and the overwhelming straight ballot pro-Democrat scenario does not come about?
We highly doubt that there will be a scenario where the consensus gets only one piece of its Trifecta forecast correct. Either the blue wave exists and it will in fact smash everything in its path (which is clearly the poll-based consensus viewpoint) or it doesn’t exist.
Blue Wave?
Alternative data simply does not pick up signs of a blue wave. For a Trifecta to come about, we would likely have seen significant shifts all over the board. Left-leaning everything and anything would be surging over the last month or so as well as going into Election Day, and we just have not seen this.
Trump and Republicans have been severely hurt by Covid and all the things that come along with it (recession, unemployment, street unrest, etc.). Prior to Covid, around February, alternative data showed the possibility of a ‘red wave’. Now, in the midst of the pandemic, alternative data show a fiercely divided country – within a larger rightward shift. It does not show a blue wave, however.
Senate Forecast
Returning to our forecast, we see mixed results. Republicans should lose seats in the Senate but could barely hold control while winning seats back in the House.
According to our analysis of Senate races, Republican candidates do not generally do well in relation to their Democratic peers on social media. Comparing Social Media Influence (SMI) of candidates in tightly contested Senate races which are normally high profile enough to produce sufficient data shows the relative strength of Democrats. According to our outlook, if this year were not a presidential election year, Democrats would have a high probability of winning the Senate.
Both The Economist and 538 forecast that Democrats will win control of the Senate with 52 seats, from the current 53 seat Republican majority. In the case of a Trump victory, we see there being an even 50/50 split.
The consensus has Democrats gaining 5 net seats while we see a gain of 3. This is not a huge difference and reflects a Trump win and relatively weak Republican candidates.
House Forecast
The outcome of the House could be a much greater surprise.
The consensus sees virtually no scenario in which Republicans can regain control of the House. The Economist puts the probability of Democrats retaining control at 99%. It also forecasts that Democrats will go from 232 to 243 seats in the House, an 11 seat gain.
In the absence of the consensus blue wave, we see little chance of Democrats gaining seats. In fact, it would be more of a question of how many they would lose.
We have two basic scenarios, neither of which involves a blue wave simply because that is not what we see in our alternative data universe. The first scenario envisions a mixed and tightly contested national election with high turnout on both sides. The second envisions a slight Republican advantage due to a Trump win.
In the first scenario, Republicans gain a net 13 seats.
Admittedly, this could be our most out-of-consensus forecast, even more so than a Trump win. However, much of the Democratic surge of 2018 was produced by unusually high turnout in a mid-term election. In most cases where Democrats flipped a House seat in 2018 from Republican control, the number of votes received by the Democratic candidate was less than the number of votes received by the Republican candidate in 2016. We have serious doubts of the stickiness of these wins.
In other words, what happens in 2020 when Republicans voters return in apparent full force?
Unless there is a blue wave, Republicans will naturally win back many of the seats they lost in 2018 just due to turnout normalizing. In mid-terms, often the out-of-power party will mobilize to a much greater extent than the in-power party. This is completely normal and is what happened in 2018. However, it is also completely normal for a portion of those seats that were lost to be regained in the next general election.
Democrats gained 41 seats in the House in 2018. For them to lose 13 seats in 2020 after such a large gain is really not that big of a stretch. Again, in the absence of a blue wave, we should only expect for turnout to normalize and under such conditions Republicans should gain some seats back.
This scenario would put Republicans at 210 seats, short of the 218 needed for control.
In the second scenario (which includes down ticket Republicans benefiting from a Trump victory and Republicans slightly improving upon their 2016 performance), we see an additional 7 seats gained by Republicans. These are seats that would be difficult to see returning to Republican hands except in the case that the party in general did well.
Bizarrely, this puts Republicans 1 seat short of a majority in the House. We believe this scenario is extremely plausible looking at the numbers and assuming there is no blue wave. Such a scenario basically makes control of the House a flip of the coin as there are bound to be upsets within the hundreds of elections.
The real question, the key question, is will there be a blue wave or not? Polls and traditional model based forecasters say directly or indirectly that the blue wave is obvious. We see this in their forecasts and the probabilities they attach to them.
Alternative data states there is little evidence of a blue wave but instead of a bitterly divided country that leans towards giving the incumbent the advantage, even in a volatile environment. The House and Senate elections would likewise reflect the divided state of the nation, producing slight majorities or an even outcome.