With almost all of the votes counted in America’s Presidential election, a mystery remains – and its solution has important lessons for British politics generally and the Labour Party in particular.
The mystery is this. While polls systematically overstated Joe Biden’s support in almost all the battleground states – predicting comfortable victories in states that Biden won only narrowly – one exception stands out. The 538 website’s final prediction for Georgia, based on a clutch of late polls, was that the two-party volte would divide Biden 50.5%, Trump 49.5%. The result was Biden 50.1%, Trump 49.9%. To come within 0.4 points for each candidate counts as a polling triumph.
Why did the polls get Georgia right, while messing up everywhere else? It’s becoming clear that some staunch Trump supporters across America shared the President’s belief that polls, like the mainstream media, were crooked providers of fake news. They refused to join online polling panels, or answer a pollster’s questions over the phone. Polls tried their best to adjust their raw data to provide an accurate picture of the electorate as a whole. However, they seem to have misjudged the size of Trump’s support in an election in which the President’s vote jumped by 18%, from his winning total of 63 million in 2016, to his losing total of 74 million this month.
So were Georgia’s voters uniquely willing to tell pollsters they were backing Trump? It seems vanishingly unlikely. Indeed, as far as Trump’s total vote concerns, Georgia followed the national trend: his increase from just under 2.1 million to almost 2.5 million matched the 18% he achieved in the rest of America.
What made Georgia distinct was Biden’s support. He won 600,000 votes more than Hillary Clinton did – a jump of 32%. Compared with his national increase of 22% (from 66 million to 80 million), Biden secured a ten-point bonus in Georgia. This was worth almost 200,000 votes in a state he won by 12,000. Without that bonus, he would have fallen well short.
Which brings us back to the polls. The paradox is that the polls seem to have called Georgia right by making two mistakes rather than one. As with the rest of the country, they underestimated Trump’s support – but, unusually, they also underestimated Biden’s support. The polls’ good fortune was that the two mistakes cancelled each other out.
So: why did the polls underestimate Biden’s support? The clue is to be found in the jump in Georgia’s registered electorate from 5.4 million in 2016 to 7.2 million this year – a 30% increase in a state whose overall population has been growing by only 1% a year. Of course the polls knew the electorate had grown, but they seem to have underestimated the main reason: a sustained and effective registration drive by the state’s Democrats, led by Stacey Abrams, one of their most determined and charismatic politicians. In the end, the polls fell short among two groups: Trump supporters who hid from the pollsters, and Biden supporters who had registered in unprecedented numbers. Like a tennis player who wins a point after mis-hitting the ball, the pollsters got lucky.
It might be thought that all this holds few lessons for Britain, where registration rates are higher than in the US. However, data from the Electoral Commission warn against complacency. Overall, around 8 million people are missing from the electoral register – either missing altogether, or registered at an old address. In percentage terms, that’s a shortfall of 15%, or more than 10,000 in a typical constituency. That’s a national average: among people over 65, just 6% are missing; among people own their home outright, the figure is 9%. But among people living in social housing, the figure rises to 18%, among black and Asian adults, to 25%; among 25-34 year-olds its 26%; and among under-25s, 32%.
Guess what: the demographic groups with the highest registration rates mostly vote Conservative, while those with the lowest registration rates tend to prefer Labour. This pattern is similar in the US, as between Republicans and Democrats. Stacey Abrams is not the first politician to notice this; her achievement has been to do something about it. She refused to accept the lazy view that unregistered adults had no interest in voting. With the right – and, vitally, sustained – campaigning, many thousands of people could be persuaded and helped to register, and then to vote.
Here in Britain, Labour’s biggest challenge is to win back the “red wall” constituencies that it used to hold with big majorities but now have Conservative MPs. Persuading voters who switched to the Tories to “return home” is plainly a large part of the task. But an effective registration drive, copying what the Democrats have done in Georgia, could add to Labour’s vote by campaigning among those who are missing from the register: those in Conservative marginals who, if coaxed to join, or rejoin, the electorate could help to eject Boris Johnson from Downing Street as effectively as the new voters of Atlanta and Georgia’s other cities helped to remove Donald Trump from the White House.
British parties have often engaged highly-paid American consultants to help them fight elections. For Labour, listening to Stacey Abrams – and doing so now, not waiting until six months before the next election – could do more good than listening to all of America’s pollsters and consultants combined.
This blog was first published by The Article