US Election:
Polling update and potential results

Still too close to call, Trump gaining ground

In our recent update, we stressed how close the 2024 US presidential election is shaping up to be. For the last month or so Democrat Kamala Harris has retained a narrow national lead, but swing state polling indicates a knife-edge result. Harris’ national lead has tightened considerably over the last two weeks, with Trump now ahead in some polls and very narrowly behind in others. Some measures, which use probability models instead of simple polling averages, have Trump ahead nationally. Whatever happens, it’s clear that this election is going to be extremely close; so far markets have reacted with caution, with sideways trends in US equity markets, while in FX markets the US dollar is strong, but not exceptionally so.

“Both the 2020 and 2016 polls underestimated the vote share of the Republican candidate”

What are polls telling us about the election?

American elections are the most polled events on earth. Every day, dozens of pollsters carry out surveys of opinion, nationally and state by state, asking how potential voters intend to cast their ballot. In order to create a poll, pollsters select a demographically representative sample from the state, controlling for age, sex, ethnicity, and income group. This sample is contacted, either by mail, email, or phone, and asked if they intend to vote, and if so, how. From this data pollsters make an informed prediction about the outcome in each state. So far, these polls show the result as too close to call in all the most important swing states, but the last few weeks have shown an undeniable move in favour of Republican candidate Donald Trump, who is pulling ahead in key swing states like Georgia and Pennsylvania.

Is polling accurate?

It is worth remembering polls only look at a small number of people, normally 1,000-2,000. A few larger polls exist, with around 10,000 participants, but these suffer from problems. Larger telephone polls are carried out using robocalls – pre-recorded voice messages – which have a poorer response rate than face-to-face meetings or human calls, or by online surveys, which are biased by the people who choose to respond to them. Another phenomenon is shy voters; people who do not give an accurate assessment of their eventual vote when contacted about it. Both the 2020 and 2016 polls underestimated the vote share of the Republican candidate for this reason. Polls come with a margin of error, normally of 1-5%, and neither candidate has been able to build up a lead which exceeds this margin, leading to uncertainty about which candidate will win in November.

Recent developments

Since the beginning of October, Donald Trump has closed the gap with Kamala Harris, moving from an average polling difference of -4 or -5% down to <-1%. That figure is a moving average of recent polls, and many individual polls now have the Republican candidate slightly ahead. Most forecasters use prediction models alongside polling averages, which run multiple simulations of the election result to find out who will win in different scenarios. Several of these models show the Republican candidate is slightly more likely to win in November. However, despite his recent momentum, traders should remember that the results remain extremely close, and that both candidates still have a fair chance of winning.

Swing states

Americans do not vote for their president directly, with the president chosen by the electoral college on a state-by-state basis. Each state has a number of assigned electors, who are obliged to vote for a presidential candidate in accordance with the statewide election results. In most states, this is a legally binding obligation, however some electors have been known to vote for a candidate other than the election winner. States are assigned a number of electoral college votes based on their population, ranging from 54 in California (population 40 million) down to 3 each for the six smallest states and territories in the US: Alaska, North Dakota, South Dakota, Delaware, Washington D.C. and Vermont. Two states, Nebraska and Maine, have a more complex system where some electors are assigned based on who comes out ahead in different parts of the state, and some based on the statewide result. Overall, the result of the electoral college is to concentrate close elections on the decisions of voters in a few states.

Do all votes count?

Most Americans live in states which are consistently won by the same party. For example, California has voted Democrat in every presidential election since 1988, and New York since 1984. On the other side, Texas has voted Republican at every election since 1976, and Alaska has only ever voted once for a Democratic candidate, in 1964. A large number of coastal states in the North East or West Coast regions are safe Democrat strongholds, and many states in the inland West or Deep South consistently vote Republican. That means voter behaviour in just a few states decides US presidential elections, and polling in these states is more important than the national figure. In this election, the most important swing states are Pennsylvania, Georgia, Arizona and Michigan; all four states Donald Trump won in 2016, but lost in 2020.

How might polls change before the election

Polls move up and down, and just like price charts in trading, everyone has an opinion on how they will move next. Pollsters talk about momentum, the weight and pressure of recent moves  – recently, this seems to be favouring Trump. An important test is whether the Republican candidate can continue this momentum and move clear of Harris both nationally and in battleground states before election day. If not, and polls remain range-bound in the 50/50 area where they are currently trending, markets will go into election night without any clear idea of the overall winner. This is a recipe for market volatility.

How are markets reacting?

The second week of October was a strong one for US equity markets, with the S&P500 trending upwards from the 7th till about the 14th of that month. Since then, a sideways trend has dominated heading towards the end of the month. This initial move coincides with the inflection point in Trump vs Harris polling, where Trump began to close the tight gap between the two candidates, but the market reaction so far has been muted. It is possible traders are still apprehensive about committing fully to the Trump trade, when national and state level polling remains so close. As we have discussed on previous articles, there is also significant policy overlap between the Democratic and Republican candidates, so it is not impossible the market reaction will be similar whichever candidate wins in November. Key assets to follow will include EUR/USD, the S&P500, and crude oil.

Conclusion: volatility ahead?

When markets have plenty of time to prepare for an election result – as we saw this year in the UK general election – it normally dampens market volatility. As traders have time to rebalance portfolios and weigh up the potential policy impact of the incoming government, they are not surprised by the election day result. Portfolio rebalancing takes place before election day, instead of after, and can proceed at a more leisurely pace. By contrast, close-fought elections that go down to the final minute often lead to volatile election day markets. ADSS traders can explore how markets might react, and which markets to watch as results come in, in our dedicated election analysis section. But as November 5th looms close, what looks sure is a close result, and choppy markets heading into the end of 2024.