I will write a longer essay on this, but there were several inputs to my expectation Trump would win. Here’s one of them.
I knew the polls were overestimating support for Harris—and I knew roughly by how much. As I have been telling people, Harris needed to be well outside the margin of error going into election night to even have a chance.
In 2016, the national polling average (aggregate) from FiveThirtyEight had Clinton at 45.7 percent to Trump’s 41.8 percent, a 3.9 percent advantage. The official vote was 48.2 percent Clinton to 46.1 percent Trump, a 2.1 percent difference, roughly half of what the polling aggregate predicted.
Clinton won almost 66 million votes to Trump’s nearly 63 million votes. The difference was not enough to overcome the electoral college.
In 2020, the national polling average from FiveThirtyEight had Biden at 51.8. percent to Trump’s 43.4 percent, an 8.4 percent advantage. The official vote was 51.3 percent Biden to 46.8 percent Trump, a 4.5 percent difference. As in 2016, the aggregate polling overestimated the Democrat percentage by nearly twice as much.
Biden won just over 81 million votes to Trump’s nearly just over 74 million. Take note of this. I will come back to it in a minute.
In 2024, the national polling average from FiveThirtyEight had Harris at 48 percent to Trump’s 46.8 percent, a 1.2 percent advantage. The official vote was 47.5 percent Harris to 51 percent Trump, a 2.5 percent difference.
See the pattern? Polling consistently underestimates Trump’s support. The error is so consistent that you can predict the outcome if you account for it.
The vote is still being countered, but as it stands right now, Harris has roughly 66.5 million to Trump’s roughly 71.5 million.
Now, about that official Biden vote count. Where did 15 million Democratic voters go between 2020 and 2024?
One has to take account of phantoms, as well.
Finally, I expected Trump to win 312 electoral votes. We are still waiting to see whether that expectation holds.
freedom and reason 