Why Trump Will Lose in 2020

Trump lost the popular vote in 2016. Less talked about is how narrow was his electoral college victory. To put that in perspective, he's the 45th president and 45 presidential elections were decided by a wider electoral college margin. Since his first day in office, his net approval across the 50 states has dropped by an average of 20 points. It is hard to imagine any scenario in which such a drop doesn't reverse his narrow margin of victory into a decisive loss.


Trump's approval rating has crashed since his first day in office. His net approval [footnote 1] across all 50 states has dropped by an average of 20 points since he took office, ranging from a drop of only 10 points [2]  to a drop of 33 points [3]. There is not a single state where his approval has gone up since he took office.

[graph with latest information at Morning Consult, here.]

He won 306 electoral votes in 2016, 38 more electoral votes than he needed to become president. That was a tight margin. (To put it in perspective, Trump is the 45th president and there have been 45 presidential elections decided by wider margins.)

I'm no scientist but I can't see how a margin of victory that slim can overcome a drop in approval rating of 20 points.

The states he won that have shifted from a net positive approval to a net negative represent 125 electoral votes. [4]

If all of those states shift from Trump to his opponent, he loses by 354 to 181, a victory roughly equivalent to Obama's 2008 win. If only one-third of the states that no longer approve of Trump vote for his opponent, his opponent wins by the very thin margin that George W. Bush won by in 2000. And obviously if his drop in approval rating doesn't change a single electoral vote, he wins again by a narrow margin. That last scenario is possible but seems wildly improbable.

And of course one thing that makes this even worse for Trump is that every year more of his supporters die. In 2016, 52% of those over 65 voted for him while only 36% of those 18 to 29 did. Every year, about 2.8 million people die and about that many become old enough to vote. Every year, Trump loses about 500,000 voters, net, as younger voters less likely to vote for him replace older voters. He lost the popular vote by 3 million in 2016. Since then, the conveyor belt of aging has meant a net loss of roughly 2 million more voters.

A cautious man would predict a narrow victory for whichever Democrat runs against Trump. These numbers don't make me feel that cautious.

----------------------------------
[1] % of the people in the state who approve of him minus the % who disapprove
[2] Hawaii, where it dropped from a negative 13 t a negative 22
[3] New Mexico, where his net positive of 17 points dropped to a negative of 16
[4] The states that have dropped from a net positive to a net zero approval rating represent another 26 electoral votes. In June, Georgia and Missouri would have to flip a coin to decide whether they approved or disapproved of him.

Predicting the 2020 Election: How the Popular Vote is Driven by the Ratio of Knowledge Workers to Factory Workers

In 1972, the Democratic National Committee (DNC) set quotas for women, minorities and students. Curiously, the party that had championed labor since before FDR, did not set quotas for representation from blue-collar or factory workers. To a degree, they turned their back on this group that Democrats had so surely represented for so many decades.

Instead of focusing on blue-collar workers, they shifted their attention to white-collar workers, the new crop of employees who created and used computers and other seemingly esoteric products.

The DNC shifting their attention from blue-collar workers in the industrial economy to white-collar workers in the information economy made for visionary policy and disastrous politics.

In the five presidential elections from 1972 to 1988, Republicans won four by an average of 11.2 million votes. Democrats won only once by 1.7 million, and that was in the wake of Watergate, the worst presidential scandal in more than a century.

If you look at this graph, you can see that in 1970, manufacturing workers were 26% of the workforce. By contrast, college graduates were just 11%. If all you knew is that Republicans had a coalition that focused on factory workers and Democrats focused on knowledge workers and that factory workers outnumbered knowledge workers by more than 2 to 1, you would predict a huge victory for Republicans. And that is, indeed, what happened. In 1972, Nixon won by nearly 18 million votes. It was a landslide. In 1976, after Watergate, Carter won as a Democrat but in 1980, 1984, and 1988, Republicans won the White House.

In the five presidential elections between 1972 and 1988, Republicans won four by an average of 11.2 million votes and Democrats won one by 1.7 million.

By 1992, though, knowledge workers eclipsed factory workers. From then on, Democrats have convincingly won the popular vote.

In the seven presidential elections since 1992, Democrats have won four elections by an average of 7.1 million votes. Republicans have won three elections by an average of -133,444 votes. (Yep. That number is negative.) In those seven elections, Republicans won the popular vote only once and - like the 1976 election after Watergate - the 2004 election was in the wake of 9-11 and the invasion of Iraq. That is to say, like 1976, in 2004 voters thought something else mattered more than their economic interests. 

From a policy perspective - a question of what policies will create the most jobs and wealth - we had easily moved into an information economy by 1972. From a political perspective - a question of what will get you elected - we had moved into an information economy by 1992. 

I'm confident that Trump will not win the popular vote in 2020. I'm doubtful that he will again win the electoral vote. Too many urban residents who identify with and are a part of the global, diverse information economy dependent on knowledge workers will vote against him and not enough rural residents who identify with and are part of the national, industrial economy dependent on capital and factory workers will vote for him.

Age is Just a Number. Then Again, So Are the Votes You Need to Win

Our last four presidents were all baby boomers.

Clinton, Bush, and Trump were all born within a year of when Japan's surrender ended WWII. Fittingly, the impulsive Trump was born a mere 9.5 months after.

Obama gets lumped in with that generation by demographers but was born 16 years after WWII ended. Less boom than echo.

Clinton, Bush and Trump time traveled together, hitting all the major events at the same age.

It is not just that the 2020 candidates include what could be the first Jew, black, woman, black woman, or gay to be president. It also includes members of the silent generation, baby boomers, gen x, and a millennial. These candidates have hit major events at very different stages of life and would bring a very different perspective to them. Bernie Sanders was 4 when Japan surrendered and the horrors of World War II would have been relayed to him as fresh stories by family members when he was a little boy rather than as distant history as it would have been for Pete Buttigieg who will be younger when he's sworn in than Sanders was when Buttigieg was born.

Here is a way to think about the age of the candidates in terms of where they were when historic events happened. (And obviously for the candidates, the age when "sworn into office" assumes that they would actually win the 2020 election.)

Oh, and yes. I know that Ocasio-Cortez is too young but I thought it was interesting to put her there for comparison. And she's too young for 2020 but she'll be old enough by 4 months to run in 2024.