The result is still unknown but we can already discern important truths about polling, policy and political strategyby Peter Kellner / November 4, 2020 / Leave a comment
At the time of writing, Joe Biden is the probable but not certain winner in the US. If he does win, it will essentially be because his strength among suburban and blue-collar voters in the north overcame Donald Trump’s strength among rural and a significant section of Hispanic voters in the south. (In a few hours we should be able to judge whether Trump’s gains among Hispanic voters are general, or mainly among formerly Cuban families. If Biden wins Arizona, it would suggest that the Democrats have a narrower, Cuban, problem.)
Before we consider the lessons of the way Americans appear to have voted, a word about the polls. Taking the final averages on the FiveThirtyEight website, they understated Trump’s support, and therefore failed to predict his likely victories, in Florida, Georgia and North Carolina. Their errors were not great: on any sensible reading of the final polls they were too close to call. Nevertheless, they seem to have systematically failed to measure Trump’s true strength in the south-eastern segment of the country.
Elsewhere, the polls seem to have done better. Subject to the final figures making nonsense of what follows, they seem to have come close to the voting figures in other battleground states such as Ohio and Texas. (The polls showed Trump on average one per cent ahead in both states. An equally narrow Biden win in either or both states—and big wins look improbable—would be good for the Democrats without being bad for the pollsters.)
It is too early to judge the polls’ national predictions—an average eight per cent win for Biden. Four years ago, they expected Hillary Clinton to win the popular vote by 52-48 per cent of the two-party share; she “won” by 51-49 per cent. The polls’ problems centred on specific states, especially the rust-belt states that in the end delivered victory to Trump. It looks as if the provisional verdict on the pollsters this time is similar: not too bad overall, but disappointing in some key states—though not the same states as in 2016. It may be a case of solving one set of problems and being tripped up by another.
I leave to others the big political implications of the result, for the US and the world—not least because until we know the final outcome, any decent…