Keys to follow the final stretch of the US elections

With only two weeks to go before the elections, we highlight the evolution of the polls for the Presidency and the Senate of the country. We also refer to the differences between the 2016 and 2020 elections and explore the current risks for the favourite Joe Biden. A week is a long time in politics and we still have two weeks left.

At the moment, Biden's impressive lead remains and has even increased slightly in the last week. He now leads by 11 points in the national polls and is ahead in each state on our tracker by at least 3 points.

Meanwhile, the Senate elections continue to show promising numbers for the Democrats and are consistent with our base scenario that establishes a Democratic victory.

We believe that, in North Carolina, despite the Democratic candidate facing allegations of cheating, the scandal has not been enough to tilt the race in favor of the Republican candidate so far. Meanwhile, in Georgia a run-off election is likely to be held in January.

These trends are particularly encouraging for the Democrats, given the huge numbers of early votes we are seeing this year. Current monitoring suggests that over 30 million Americans have already voted. This participation is about 5 times higher than what has been seen so far in 2016. In fact, this represents about 22% of the total votes counted in those elections. This could reflect greater enthusiasm, pointing to a higher turnout in 2020, or a COVID effect, as more people mail in their ballots or vote early to avoid the crowds. Early turnout has been particularly high among Democrats (54%), with Republicans at 25%, although we do not have data for all states.

Of course, a week is a long time in politics and we have two weeks left before Election Day and more than 75% of the electorate have yet to vote. Naturally, comparisons are being made with this time four years ago when Hillary Clinton was expected to win the Presidency and finally, despite winning the national vote, lost the Electoral College. We believe that there are three major differences between 2016 and 2020:

1. The candidates - We believe that Hilary Clinton carried unusually great weight in the 2016 election, including gender bias, her marriage to a former Democratic president and deep scepticism among voters about her transparency. A last-minute letter from FBI Director Comey just before the election, effectively used by the Trump campaign, was crucial in driving down their polls and ultimately their loss in the election. Biden, on the other hand, has some political baggage, but it does not seem to weigh as much and the Trump campaign's attempts to label him as corrupt have largely failed.  That said, we believe that Biden is probably even more vulnerable to unexpected events than Trump.

2. The context - When Trump ran for President in 2016, he was a relatively unknown politician, but a high-profile media figure. In 2020, voters know the president well and have seen his political approach, particularly during the recent COVID-19 crisis. Over 50% of Biden's voters cite "he's not Trump" as the main factor in going to vote. Furthermore, certainty levels are high so the effect of undecided voters is probably small.

3. The polls - Biden is 4 points above the level Clinton was at this time in 2016, having enjoyed a more stable campaign over the summer and even increased his lead in recent weeks. This also means he has more room to lose support in the polls while maintaining a healthy lead. Polls in 2016 struggled to capture the impact of education on voter behaviour, which many pollsters are now adjusting. This may be positive for the accuracy of the polls, although we will not know until we have the results of the election. There is always a risk that pollsters adjust their methodology too much or too little, or end up fighting the final battle when the dynamics that drive elections have changed. There is also a risk that Trump voters are escaping the radar of pollsters after an unusually bad news cycle for the Presidency, which could exaggerate Biden's advantage. Fortunately for Biden, an eleven-point lead provides a good cushion for poll error.

We should not be complacent about the result despite the differences from 2016 that seem to benefit Biden. US elections are unusually difficult to predict because the outcome depends on the electoral dynamics in fifty states, which can vary greatly in terms of rules, regulations, voter suppression, registration and turnout.

However, in our opinion, Biden has a 75% chance of winning the Presidency. We will be keeping a close eye on the risk of scandal, the attempts by the Trump campaign to discredit Biden.

We continue to give Trump a 25% chance of re-election in two weeks time, reflecting the high levels of uncertainty discussed above regarding the dynamics of individual states and winning voter support. For example, Republican officials point to strong Republican voter registration in key states such as Florida and Pennsylvania.