The United States goes to the polls in just on a fortnight. I (and others) have been saying for a couple of months now that Democrat Joe Biden had a sufficiently large and stable lead in the polls that he would be expected to win in the normal course of events – in other words, that if Republican incumbent Donald Trump was going to hold on, something was going to have to happen to shake the race up.
The time for that something to happen is now very short, so the odds against it are correspondingly high. Biden’s lead has not only failed to narrow, but has actually widened: for the last week and a half it has been above ten points, and this morning is sitting on 10.3. That represents a two-party swing of 4.4%, which on a uniform basis would net the Democrats an additional eight states and a majority of 198 in the electoral college.
Not surprisingly, FiveThirtyEight’s forecast model has had Biden tracking steadily upwards and now gives him an 87% chance of victory. That’s up from 77% in the last month.
But not everyone agrees. Back in September I pointed out the way that the betting market had diverged from what both the polls and the experts were saying. That divergence disappeared in the wake of the first televised debate; now it has reappeared.
Maxim Lott and John Stossel’s aggregation of betting odds gives Biden only a 61.2% chance; that’s fallen from 65.4% in just the last nine days. Sportsbet has him at 7-4 on, for an implied probability of 59.3% – extraordinarily generous odds for someone with a ten point lead. A couple of days ago it went as low as 57.1%. Perhaps Trump supporters have more money to risk, or are more resistant to looking at the evidence.
This time, the state-by-state story is much the same. Lott and Stossel now give Biden an edge of only 72 seats in the electoral college; Trump is now favored (very narrowly) in Florida, the largest of the really marginal states, despite the fact that he has trailed in the polls there for the last six months. Biden also leads in the polls in neighboring Georgia, but the betting market puts his chance there at less than 40%.
Now, it’s true that Biden’s lead in key states is not quite as big as you might think from the national polls. If the swing was uniform he’d have a lead of about eight points in Florida, whereas at best he’s getting half that. So it may be that rather too many of the votes flocking to Biden nationally are in states that are already safe for one party or the other.
But the effect isn’t large; none of the other marginals are underperforming as badly as Florida, and the occasional one (like Iowa) is overperforming. It’s not enough, as I see it, to justify the discrepancy between the polls and the odds. More likely that discrepancy will iron itself out in the coming days, just as it did last time.
And the good news for Biden is he doesn’t have to win Florida or Georgia, or even Arizona and North Carolina, where he’s still favorite (although in the latter only just). He can make no gains at all in the south and still win, as long as he picks up Michigan, Pennsylvania and Wisconsin – in all of which he has very comfortable leads.
Of course, Biden wants more than a narrow win: given the president’s attitude to democratic process, he wants a margin that’s decisive enough to be beyond dispute. (And to win a Senate majority, which we’ll talk about in a separate post.) Whether he gets it will depend on what happens in the next fortnight.