Is the US election a done deal? Not yet, says Dick Roche, and explains why it is too early to consider that Donald Trump will necessarily lose.
Dick Roche is a former Irish minister for the environment, heritage & local government, the department that is responsible for elections in Ireland.
Four weeks out from polling day and with American voters already casting their votes in early voting or by mail Joe Biden is polling ahead of Donald Trump in key battleground states. It is hard to see how Trump can turn things around.
However, 4 weeks out from the 2016 elections polls favoured Clinton and forecasters put her well ahead of Trump. Three days before the election the Princeton Election Consortium found that Hilary Clinton had a 99% chance of election. The New York Times said she had an 85% chance of winning.
In the sense that Mrs Clinton received almost 2.9 million votes more than Mr Trump the national polls were not wrong- but that’s not how US elections work. Winning the national popular vote does not translate into a win in the Electoral College which picks the President and Vice President.
When the Electoral College came to vote on December 2016, 304 votes went to Trump, 227 votes went to Clinton. Seven Electoral College Electors went ‘rogue’ and voted for neither candidates.
In 2016 Trump won Michigan, Wisconsin and Pennsylvania, less than 1 point. Those three states gave him 46 Electoral College votes and the keys to the White House. As of 10th October Biden is ahead by between 5.5 and 7.1 points ahead in those three states.
In five other states that Trump won in 2026 Iowa, Ohio, North Carolina, Florida and Arizona Biden is registering leads ranging from 0.6 to 3.7 points.
Biden is also recording more positive poll figures in states which Clinton won by slim majorities in the last presidential election and is performing very well in the traditional Democrat strongholds.
So is the election a done deal? The answer is not yet.
Polls can also be thrown off course by factors that distort their findings. In 2016 a number of factors tipped the US polls in the wrong direction.
A higher than expected turnout of voters in the lower socio-economic levels, a cohort that tends to be reluctant to respond to pollsters, impacted on polling predictions in 2016.
“Shy Trumpers”, people who were unwilling to be frank about their voting intentions because of concerns that they could be ridiculed for their views was another factor.
Hostility towards polling also played a part: people distrustful of media were unwilling to cooperate with pollsters and again impacting on the accuracy of the polling.
The same factors could shave points off the Biden lead in key states this time around. The latter two factors are if anything stronger in 2020. These could be suppressing Trump’s numbers in the polls.
Another factor that caused problems in 2016 was gauging voter commitment. Post-election analysis showed that Clinton supporters stayed at home. Polling by the Pew Research Centre in August found that 56% of those supporting Biden were doing so because “he is not Trump”.
Determined voters cast their votes come hell or high water and there are a lot of determined voters in the US this year. Whether the raw enthusiasm Mr Trump generates amongst his base is a greater motivator than the loathing he generates will be tested in November.
The biggest difference between 2020 and 2016 is the circumstances in which the election is taking place. As of 4th October 7.5 million COVID 19 cases and 210,000 related deaths had been recorded in the US. No election has ever been held in similar circumstances. The pandemic is a key campaign issue.
How the pandemic impacts on the electoral process is a separate matter. The 2020 US Presidential election will be unlike any previous election. Up to 75% of US voters have the right to vote by mail (VBM) and an unprecedented number will exercise that right.
Setting aside the arguments for and against VBM, the handling and processing of tens of millions of in mail ballots will impact on the election outcome and do so in a way that is very difficult for polls to capture.
The process of in-mail voting is significantly more complex than in-person voting. It requires detailed planning, skilful administration, involves more actions by the voter, by election officials and third parties. As a consequence, there are more occasions for spoiled or rejected votes with in-mail voting than with in-person voting.
In the 2016 US Presidential elections 23.7 % of all votes were cast by mail, almost 1%, over 318,000 of those votes were rejected. In the US 2018 mid-terms, the rejection rate rose to over 1.4% and 430,000 plus votes were rejected.
An analysis by National Public Radio of the 2020 primaries across 29 states and New York City put the number of rejected in-mail ballots at over 550,000.
Strategists on both sides of the political divide predict a record voter turnout in this year’s presidential election. In July 41% of voters indicated they were “very likely” to use in-mail ballots with another 23% “somewhat likely” to do so.
A study published in August found that 52% of voters were likely to vote by mail in the election. That figure included 72% of Democrat voters but only 22% of Republicans.
If only 41% of those voting in the presidential election use VBM and the rejection rate remains at the 2018 level over 1 million in mail votes could be rejected. Projections suggest very high levels of rejected votes in key battleground states that are using VBM at scale for the first time.
The impact of 1 million “lost votes” be heaviest on the Democrats both because its supporters are more disposed to use VBM and because key cohorts of Democrat voters have been shown to be more likely to have their in mail votes rejected because of voter error.
The factors that undermined Clinton’s polling advantage in 2016 still exist and could cut back Biden’s poll lead. Rejected VBM ballots could shave further points off the Biden lead in key states making the election closer than it seems.
How the pollsters can factor high levels of lost votes into their work is not clear. What is certain however is that if there is no clear-cut winner on election night, rejected VBM ballots will become the key focus.
In battleground states, every vote cast by mail will be closely scrutinised and every error will be challenged, automatic recounts could be triggered in many states, multiple bitter court cases could follow and, as in 2000, the presidential election could be determined by a Supreme Court decision.
It would be the ultimate irony if the vote by mail which has been so heavily promoted by the Democrats were to become the last available lifeline for Trump.