Joe Bid leads the Real Clar Politics average by 7.1 points and the FiveThirtyEight polling average by 7.5 points. He has a clear lead in almost every true swing state. In Nate Silver's model only one state is categorized as a "toss-up", which means neither candidate has a greater than 55% chance of winning. That state is North Carolina. Mr. Silver's forecast has Joe Biden winning North Carolina 53% of the time. The second closest race is Ohio, which Trump is currently winning in 57% of simulations.
A lot of the concern about Biden's momentum was based on a stutter-step in the 538 model. When Trump got the slightly more favorable post-convention bounces, Biden's chances of winning dipped below 70% for a time, which changed the characterization of the race from "Biden Favored to Win" to "Biden Slightly Favored to Win."
A few days ago Nate tweeted out an important explanation of how his model works. His mode factors time remaining quite heavily. Polls a few months out are subject to a lot of change, and Nate balances those out by considering previous voting behavior and partisan fundamentals in each state. If you click on his Wisconsin forecast, you learn that "Right now, 56% of the forecast for North Carolina is based on polls and 44% relies on factors like demographics and past voting patterns." As the election gets closer, the polls will count more and the other factors less. Silver summarized this effect by saying that if the election were being held that day, Biden would be about a 90% favorite to win.
So the difference between the Economist's bullish forecast and Nate's more guarded forecast is time. As a general principal, I tend to favor Nate's logic. But you do have to consider the specifics of this election. A lot can go wrong in eight weeks, but Trump's biggest obstacle seems to be that the American people have firmly fixed views of him, and they tend to be negative.
Trump will get over 40% of the vote. He might get over 45%. Because the electoral college structurally favors Republicans, that might be enough for him to win. But so far he has tried several approaches to attacking Biden and none of them have worked. I simply do not think that he's going to be able to make up for his three years and eights months of shitty, incompetent, and corrupt governance. Trump will fight hard and he will fight dirty. But he is a the under dog, and for very good reason.
The only real news this week was that Trump's campaign may be having financial troubles. Biden raised $364 million dollars in August. That is more than Hillary Clinton and Donald Trump combined to raise in August 2016. Trump has not yet released his number for August.
I think some of the reporting on his cash woes is a bit overstated. I do not think he will have any trouble being on the air waves in the closing weeks of the election. But big Republican donors can read polls too. If they think Trump can't win, they may start diverting funds to important Senate races.
I revise my forecast slightly by giving Biden an 85% chance of winning. That is up from 83% last week and back to where I had it 3 weeks ago. My prediction remains Biden 334, Trump 204 in the Electoral College. Nothing I have seen recently has convinced me that a single state has changed.
I think Biden will sweep the Big 3 Rust Belt states that swung the election last time (WI, MI and PA). I also think he will win Arizona and Nebraska's 2nd congressional district. I remain cautiously optimistic that he will win Florida and I favor him slightly in North Carolina.
The states to watch for an expanded map are, in order, Georgia, Iowa, Ohio and Texas. For now, I assume Trump will sweep them.
No comments:
Post a Comment