With both party conventions now in the books, it is a good time to look at where we are in terms of the election, and what is likely to happen. Luckily, Decision Desk HQ and 0ptimus have run back the successful 2018 House and Senate models, with a new Presidential one as well, and therefore there are many interesting takeaways from the modelling. This is not an exhaustive list, but here are six interesting and notable takeaways from the DDHQ/0ptimus model.
A Biden Landslide Is Possible, But Unlikely
The model gives Joe Biden a 85% chance of winning the election, on the strength of winning the Midwest and Florida, being tied in Ohio, North Carolina, and Arizona, and only being narrowly behind in Texas. They don’t treat Biden’s victory as a foregone conclusion, either, but they don’t dismiss the notion of a landslide either. The multiple paths to 270 are also reflected, as the state by state odds are lower than the cumulative probability of a Biden win, meaning they understand that there can be many, many ways for Biden to win. The model doesn’t overrate Biden either – situated between Nate Silver’s model thinking the race is 70/30 and The Economist’s thinking it’s 90/10, the model shows clearly how Trump could win again without either overstating his position or dismissing it.
Not All House Incumbents Are The Same
One of the good things about the model is that it doesn’t treat all House incumbents as equally valuable to the prospects of winning their seats. Two GOP incumbents got primaried – Scott Tipton in Colorado 3rd and Steve King in Iowa 4th – and the model treats those losses very differently. In the Iowa 4th, the loss of King helps the GOP, as his offensive statements and inflammatory past are now gone, and the GOP return to winning it massively according to the model (and all the rating agencies). In Colorado 3rd, meanwhile, Scott Tipton’s replacement with a candidate who has attracted controversy for ties to QAnon, and the model has her only winning this GOP held seat 23.5% of the time. Good models have dynamic variables to ensure they react to real world realities, and here the model excels.
The House Is Safe Democratic
The model only has one Democratic seat in the column of the GOP (MN-07), although some of them are Democratic leaning tossups. The 1.9% probability of a GOP house is properly under the probability of a Trump win, and the right claim is that even if Trump pulls back his lead, the House GOP are in trouble. When you have seats like the Indiana 5th in Dem hands (albeit as a Dem leaning tossup), the House isn’t flipping. Even where the model disagrees with others (like LeanTossup), it isn’t in a systematic way – in some places, they’re higher on the GOP, but in others, it is to the GOP’s detriment.
The Model Thinks Doug Jones Has A Chance
This one is a big shock – Doug Jones still has a 25% chance to win his Alabama Senate seat won when he ran against Roy Moore three years ago. Written off by many, this race is apparently closer than many (including this author) think it is. It could just be a function of some decent, albeit Democratic tilting, polls right now that could change, but this still being in play would make a real change to Democratic chances in the Senate. If they can win Alabama again, then they need the core four gains (CO, AZ, ME, and NC) to get them to 51 seats, not 50 – and lock them into the Senate majority, whatever the Presidential election does.
South Carolina Senate Is In Real Trouble
This race has tantalized Democrats looking to get rid of one of their least favourite Senators, but most people don’t view it as all that winnable. The model thinks that the GOP are only 70% to hold the seat, with the war chest and good polls of Jamie Harrison meaning that Lindsay Graham is in trouble. Democrats should worry about the chances that this is a money pit, but the chances of Harrison activating black turnout to a super high degree and whites like those who voted for Joe Cunningham in 2018 voting blue all along the ticket makes this race close.
Minnesota Is Less Democratic Than Michigan
A favorite bugaboo of mine is the chance that Minnesota, which only barely voted for Hillary Clinton, may end up to the right of Michigan this cycle, which obviously slipped into Republican hands in 2016. It is by the barest of margins, but the model does reflect the fact that the whiter, less urbanized state is the one where Biden is less likely to win, and can pick up on that sort of demographic trend. Minnesota is trending right fairly fast, and just because 2018 had Amy Klobuchar at the top of the ticket and a very blue environment, the state isn’t immune from the trend sweeping it right, and this model sees it and knows what it means.
Evan Scrimshaw (@EScrimshaw) is Managing Editor and Head Of Content at LeanTossup.ca and a contributor to Decision Desk HQ.