With Two Months to Go, What if the Election Were Held Today?
Happy Labor Day! Hope you have been able to relax and unwind. How am I? Don’t worry about that. Ignore all my vague gestures on Twitter about this website being one of the few positive consequences of a tiny little months-long mental health episode. Haha, it’s fine!
Not a whole lot new to report on the model update front. Democrats saw a small uptick in their odds across the three forecasts in the wake of their convention. Harris received some very good national polling this week while the swing state picture looks less rosy. A disproportionate amount of those state polls, however, came from fly-by-night partisan firms like Trafalgar and InsiderAdvantage.
The model, by design, doesn’t put a lot of weight on polls from sources that do not publish their crosstabs or simply pass my own subjective vibe test, but they still have a significant impact. Including them would have made the forecast more accurate in 2016 and 2020, though they were the reason the precursor model to Solid Purple ended up showing Dr. Oz being a narrow favorite in the 2022 PA-Sen race. If I had just ignored Republican pollsters flooding the zone and stuck to the rivers and lakes that I’m used to, it would have been Lean Fetterman.
We’re all just doing the best we can here.
While Harris now has a 56% chance of winning the election in our forecast, Nate Silver now has Trump back in the lead, despite being down in the popular vote almost as much as he was when he lost in 2020. Never mind that he was shitting all over G. Elliott Morris and FiveThirtyEight for being bullish on Biden back in July because the “fundamentals” painted a much rosier picture for his re-election than his polls. Silver’s model prices in an expected “convention bounce” that, because Harris isn’t winning by 6-7 points right now, must mean that the election is actually much closer than it appears (and it’s still pretty effing close!)
Solid Purple handles the added uncertainty in late summer polling in a much more simplistic way - by allowing for a wider range of potential outcomes, making each race appear closer than it would otherwise. This may lead you to ask: what if the election were held today? After all, we’re only a few days away from early and mail-in voting starting in states like North Carolina, so in some respects, election day really is about to occur.
A few methodological points: between now and November, the forecast will rely much more on surveys of individual states, especially in the Congressional races. It still relies quite heavily on the national environment, since races like OH-Sen and MT-Sen historically look close or winnable for the intuitive underdog until voters tune in and sort into their natural ideological constituencies. It’s also much more confident. The model is calling Iowa “Lean R” for Trump right now because a 7-8 pt. lead during convention season is historically flimsy, but would I advise you to put money down on Harris having a prayer there? Nah, not really.
So without further ado, here is what Solid Purple would be forecasting if today was Election Day:
President
Kamala Harris would have a 62% chance of winning the Electoral College, while Donald Trump would have a 38% chance. Harris would be projected to win the national popular vote by 3.3%, identical to her projected margin in the regular forecast.
Given how close to 50/50 things are currently, there is not a huge difference between the probabilities in each individual state, or the odds of various scenarios playing out. The most notable difference is that a “nowcast” finds it significantly more likely that Trump loses the popular vote but wins the election anyway, since the odds of various black swan events (Blexas!) decrease significantly as time runs off the clock.
Senate
The Senate forecast, presently giving Democrats a 36% chance of holding onto the chamber despite an enormous uphill battle to hold onto seats in Ohio and Montana, and if they can’t, making up for it somewhere like Texas, Florida, or Nebraska all of a sudden. That may be a touch too rosy. The nowcast puts their odds at a slightly lower 31%, because while they would be substantially more likely to win in places like Arizona, Michigan, Nevada, Pennsylvania and Wisconsin, they still have to thread the needle of pulling out an upset in at least one race they are not likely to win.
House
The House of Representatives literally cannot be closer at this point, so regardless of how much uncertainty you factor into the equation, a coinflip remains a coinflip. The 435 House districts are sorted as neatly as you could possibly imagine, and while it is quite possible one party wins all the tossups and builds a robust majority, there simply isn’t any way to know until it happens, assuming polls do not shift decisively in one direction.
Nine more weeks to go, y’all. Be kind to each other and yourself. Will I be taking this advice to heart? Shit no.
