Please remember this is done in fun because I love doing this type of work as it hits right at my natural sense of curiosity and competitiveness. This is not meant to make a political statement or try to push my beliefs on to anyone. My job is to predict what may happen in the future and the potential outcomes for financial markets. 

I would also like to thank Alex Drew for helping me compile this data.

Over the past few weeks, we have been looking at different data points, dissecting the election to predict who will win. I am sticking my neck out once again and predicting that Donald Trump will not only win the election, but he will win by a wider margin than he did in 2016.  I have Trump winning the electoral college with 315 votes, to Biden’s 222. I think Trump will LOSE the popular vote, probably by a wider margin than he did the first time.  

I went through all of this in a video early this week and what it is likely to mean for the stock market. 

(My Map Created Using RCP)

Donor Data

The first point I focused on was how many people were donating to each candidate. Based on that donor data, Trump is likely to carry the same swing states he won in 2016. 

Polling

Second, after reading through many of the swing state polls, I found that many were making the same mistake as in 2016. More importantly, many polls show a spread with Biden beating Trump that falls within the margin of error. For example, several Florida polls show Biden winning, like this NBC/Marist poll. However, that NBC poll has a margin of 4.4% error, placing the difference between the two candidates in that margin of error, basically a statistical tie.

(Real Clear Politics)

Polls in Georgia also show that all of the outcomes fall within the margin of error. In North Carolina, 10 out of 11 polls fall within the margin of error. There are 6 polls in Texas, with 3 falling within the margin of error, and Trump winning by 4 points or more in the three polls outside the margin of error. 

(Real Clear Politics)

Social Desirability Bias

More interesting are the questions some of the polls asking and my understanding of the Social Desirability Bias. When the respondent to the poll tells the pollster the answer, they think the pollster wants to hear or doesn’t give the pollster their true opinion. 

For example, this Michigan poll shows Biden leading by 10 points on October 25. However, on October 18, the party sample was broken down with 39% of Democrat party participants, 39% Republican, and 18% Independent. It asked the participants who they thought their neighbors would vote for. The results showed that 39% thought their neighbors would vote for Trump, 36% for Biden, and 25% were not sure. Then on October 25, another poll was conducted, and it had a party breakdown with 43% Democrats, 39% Republican, and 13% Independent. But this time, when asked how their neighbors would vote, 43% said they would vote for Biden, 41% for Trump, and 16% were not sure. Point being, I doubt that Biden is winning by 10 points in Michigan. 

In a Wisconsin poll, Biden had a 4 point lead over Trump among registered voters 47-43. Interestingly, 47% of the respondents thought Trump was doing a good job, up from 44% in August. I would question why someone would vote Trump out of office if they thought he was doing a good job. Again, it just seems odd to me. 

Demographic Shifts

In 2016, Trump won Pennsylvania by about 68,000 votes. But more interesting is that in 2016 there were 4,217,456 registered Democrats, 3,301,182 Republicans, and 1,204,339 Independents in Pennsylvania. In 2020, that number shifted to 4,229,158 Democrats, 3,542,265 Republicans, and 1,318,575 Independents. The Republicans gained an extra 241,083 voters, while Independents gained 183,145, and the Democrats gained 11,702.

Even if all of the new independents swung to Biden, the Republicans would still have a net gain of about 115,000 new voters. On top of that, Trump would need to lose the 68,000 vote advantage he had from 2016. It means Biden needs to gain around 183,145 votes, or Trump would need to lose them.  Additionally, Biden probably needs to get around 3 million votes to beat Trump; Obama had 2.9 million in 2012. Not easy.

Florida is not all that different, with the number of Republican voters rising by 475k since 2017, while the Democrats grew by 395,000. It gives the Republicans a net 80,000 voter advantage over the Democrats. Trump won Florida in 2016 by 119,770 votes. 

Meanwhile, North Carolina is the same, with 2,736,124 Democrats in 2016 and 2,099,551 Republicans. In 2020, there are 2,607,337 Democrats and 2,208,384 Republicans, which means the Republicans have gained a net of 237,000 voters.  Trump won North Carolina in 2016 by 177k votes. 

Other Minor Factors

In Iowa, Biden came in fourth place in the caucus behind Pete Buttigieg, Bernie Sander, and Elizabeth Warren. 

Arizona is really close; at this point, with just one poll outside of the margin of error and it has Trump up by 4 points. 

(Real Clear Politics)

Overall, this is likely to be a very close race, and we must remember that the national poll and popular vote do not matter. The only thing that matters is the state vote, and while many people may be turning out this election, if they all turn out in California and New York, it doesn’t help Biden. Each state only has so many electoral votes. The system was specifically designed to prevent one state or one part of the population from determining the winner. 

The good news is that people are voting this year, and I am rather confident that no matter who wins, the country will endure and move forward as it has for the past 235 years. 

-Mike