Following the debate between Kamala Harris and Donald Trump, the Democratic candidate has taken a step forward in the polls and among forecasters. But beware, it is a small step: the election is still extremely close.
Our average of predictions, taken from seven reliable and complementary sources, says Harris has about a 56% chance of winning, compared to 44% for the Republican. It’s almost a statistical tie: imagine that we have 20 possible futures ahead of us: Harris would win in 11 and Trump in 9.
All forecasts are similar. Statistical models of The Economist and FiveThirtyEight raise Harris’s chances to 60%, while Nate Silver’s model and forecasters Polymarket They see a tie. It’s not a very substantial difference, but Silver’s numbers surely deserve special attention because he has the better track record.
The situation in key states
As we explained, the election actually depends on the result in each state. To win the presidency, candidates need 270 delegates, or electoral votes, adding up those allocated by each territory, which all go to the candidate who wins the district.
Knowing what’s happening outside means understanding what’s going to happen inside, so don’t miss anything.
KEEP READING
Right now, Harris has 225 likely or certain electoral votes, almost the same as Trump, who has 219. These are the votes from the territories where they are favored, but we should not take them all for granted either. For example, Virginia and Minnesota are assigned to Harris, because she has a 90% chance of winning there, and Texas and Florida are assigned to Trump for the same reason, but they are places where surprises could occur: events that have a 10% chance of happening, it is more or less as easy as footballer Erling Haaland scoring three goals in a game.
That said, the main source of uncertainty remains the so-called key or hinge states, marked in grey on the map.
The key states are those with the most divided electorates, those that can easily change hands: Michigan, Wisconsin, Pennsylvania, Nevada, Georgia, Arizona and North Carolina. The last two presidents, Trump and Joe Biden, won six of these seven territories.
What is the situation now? The following table summarises the probability that Harris and Trump have of winning in each state, according to three sources: The Economist,FiveThirtyEightandMetaculus.
These state data explain why Harris is slightly ahead of Trump. Right now, the Democratic candidate is favored to win Michigan, Wisconsin, Pennsylvania and Nevada, which would probably give her 275 electoral votes and the presidency.
But the table also makes it clear that her lead is short. First, because what those predictions say is that she could well lose in those territories, with probabilities between 35% and 47%, depending on the case. In addition, Harris needs them all. If she loses any, she would have to make up for it with Arizona, Georgia or North Carolina, for example, which a priori are more difficult.
In the coming weeks, we will talk a lot about possible combinations of states won and lost by Harris and Trump. But it is important to understand one reality: the result in each state is not independent of the others.
That is, we cannot think of Michigan, Wisconsin, or Pennsylvania as separate coin tosses. They are connected. If a candidate surprises and does well in one territory, he or she will likely do well in much of the country, especially in places that are more socially and geographically similar. The correlation between states will not be complete, because each place has particularities, but as a mental shortcut, it is more accurate to imagine the elections as a national toss-up than as dozens of regional toss-ups. One example is the 2020 swing: when the country moved toward Biden, five of the six key states suddenly swung.
That’s why the equality we see now is compatible with seeing Trump or Harris win by more than 300 electoral votes on Election Day. If a candidate does better than expected, and beats the national polls by two or three points, that can translate into many states changing hands.
