October 10, 2020

STOCKS – NONE

Macro – NONE

Mike’s Reading The Markets (RTM) Premium Content – NOW WITH A 2 WEEK FREE TRIAL 

• Trump And Biden May Actually Be Tied
• Stimulus Hopes Will Boost Stocks Higher Into The Election
• Short-Term The Sell-Off May Be Finished, With ATH’s Coming
• Stocks May Retest September Lows
• Cisco May Surge Based On Options Bets
• Biden’s Lead Over Trump May Not Be As Big As It Seems
• How To Find Technical Trends And Levels

Many of the polls currently show that President Donald Trump is losing by a relatively wide margin to the challenger, former VP Joe Biden, for the Presidential election. The latest average from the from RealClearPolitics shows that Joe Biden has a 9.6 percentage point lead over Trump.
However, when running through some data and looking at things from a different perspective, we can find that the race may be much closer than those poles would suggest.

One project that I worked on starting in 2016, which proved to be a unique approach, was to use the Federal Election Commission donor data. I took the concept that a donation towards a candidate was the same as placing a bet or wager. As part of this process, I only looked for donations of $5 or more, and I removed all individual that donated more than once, so we all only get one vote.

In the 2016 election, in the swing states, if Donald Trump captured at least 55% of the total donations, he won that swing state. When I ran through the data for 2020, my results showed him with more contributions in many of those swing states, and that Joe Biden, in most cases, had more donors than Clinton in 2016. The data also shows that Trump has more than 55% of the total donations in each of the swing states, except for Pennsylvania.

I understand there may be potential flaws in this method. The exercise is meant to help determine if a candidate has more or less support than the previous cycle. I also believe that if someone is willing to donate to a candidate, they are more likely to vote for that candidate.

We care about this type of data because, historically, on average, people vote their party line. On average, since 1992, typically, Democrats represent 38.5% of voter turnout, while Republicans represent about 34.3%, and Independents represent 26 1/2%. Digging deeper, one would find that, on average, Democrats will vote for the Democratic nominee 87% of the time. Meanwhile, Republicans will also vote for the Republican nominee 87% of the time. The Independent vote has gone 45% of the time to the Democrat nominee and 44% to the Republican nominee.

What made the election in 2016 very different was that Donald Trump was able to capture a large amount of his base, and the independent vote – 90% and 47%, respectively.  

This is why understanding a poll’s breakdown and cross table are fundamental in understanding the outcome of that poll. For example, in the latest, The Economist/YouGov poll, the number of Democrats far outweighed the number of Republicans. For example, there were 1,495 participants with 562 Democrats (37.5%), 523 independents (35%), 410 Republicans (27.5%).

However, in 2016, the voter turn out was very different, with 37% of the voters representing Democrats, 33% representing Republicans, and 30% of the voters representing independents. Also, we find 90% of Republicans voted for Trump, while 89% of Democrats went for Clinton. Trump managed to capture 47% of the independent vote, while 42% went to Clinton.

These minor facts are important because when we look at the latest poll in the example above, we find that 96% of the Democrats would vote for Joe Biden. In comparison, 96% of the Republicans would vote for Donald Trump. However, since the population’s breakdown in the poll is so much different in terms of what real voter turn out is, the number of votes going towards Donald Trump is being undercounted. Meanwhile, the poll shows 47% of the independent votes are going towards Trump, and 38% are going to Biden.
If we were to re-randomize the 1495 people towards something that resembled 2016 using the same cross tables, we would find a very different result. Based on the same cross-tabulations that 716 people would favor Biden. In comparison, 695 people would favor Trump, which means that the percentage would have been 47.8% for Biden and 46.5% for Trump, 5.7% of the votes going to no one.

Again, This is why sampling and understand the break down in terms of the polls are necessary. It would also suggest that the percentage lead may not be as wide as it may appear, which means that there is a very good chance there is no blue wave in November.

As you begin to look through more data and other polls, take the extra step, you may be surprised by the outcome.