Trump is losing
Welcome back to Bournbrook’s regular series on the state of the US presidential race, your impartial guide to the campaign and what you can expect as election date draws near.
On Friday morning, President Trump announced that he has tested positive for the coronavirus, alongside his wife, Melania. The president happens to fall in at least two at-risk categories, being seventy four and overweight. So far, he is reportedly showing mild cold-like symptoms, having recently been admitted to hospital ‘out of an abundance of caution’. Trump will, of course, be receiving the highest quality care.
How exactly this might impact the campaign, if at all, is far from clear. On the one hand, it is possible that Trump might benefit from a so-called ‘rally round the flag’ effect, although I remain sceptical. Contrary to popular belief, when Boris Johnson fell ill from covid in March, it did not result in an uptick to his approval rating (which began days before the PM tested positive). Moreover, perceptions of the president are almost wholly locked in, so any meaningful sympathy effect remains unlikely.
On the other hand, there are a few potential dangers for Trump (besides the obvious health concerns that come with contracting a deadly virus). The revelations will likely shift the news agenda back towards the pandemic, a scenario which the Trump campaign precisely wanted to avoid in the final weeks. Trump’s rallies, one of his most proven methods for generating unfettered television coverage, have now been cancelled. Instead, should his condition deteriorate, news anchors might begin speculating whether he is even physically fit to lead, a criticism which Republicans have frequently levelled at Joe Biden.
Voters may further question the president’s prior comments where he downplayed the virus. As recently as last week, Trump had claimed that covid ‘affects virtually nobody’. They might also recall how, just days ago, he had mocked his opponent for frequently wearing a mask. Speaking of which...
On Tuesday, Joe Biden and Donald Trump faced off in the first of three presidential debates (whether the rest will even happen is now unclear). It was, for the most part, a trainwreck, with Trump attempting to throw his challenger off-balance through constant interruption. The result was hardly Biden’s best debate performance, as the Democrat often appeared flummoxed and unable to complete his train of thought. However, the president’s aggressive tactics likely did more harm than good.
To be clear, this isn’t my opinion. Republican pollster Frank Luntz hosted a virtual focus group of fifteen undecided voters, and few had anything good to say about Donald Trump’s performance. The president was described as ‘unhinged’, ‘arrogant’, and ‘chaotic’. Biden, in turn, was thought to be ‘coherent’, ‘presidential’ and ‘better than expected’.
In the months leading up to the event, Trump and his campaign have scored an immense own-goal by lowering expectations for their opponent, whom they claimed was ‘senile’ and suffering from dementia. According to a third of the focus group, the president’s worst moment came as he seemingly declined to condemn white supremacists. It should have been an easy question, yet it ultimately ended with a number of senior Republicans distancing themselves from their nominee.
In terms of a more quantitative approach, scientific polls conducted after the debate fared no better for Donald Trump. A CBS/YouGov poll found that voters thought Biden won by a seven-point margin, almost identical to his current national lead. Morning Consult found that voters favoured Biden’s performance by a much larger 16%. In another survey, conducted by Ipsos for FiveThirtyEight, 59.7% rated Biden’s performance as ‘good’, compared to 39.2% for Trump.
Sidenote: emphasis here on ‘scientific’. Yes, President Trump won a bunch of online ‘clicker’ polls, which are hardly more representative than a campaign rally. If the last few years have taught us anything, it’s that Twitter is not real life.
I should repeat what I wrote about the debates last week - it is highly doubtful that they will have a meaningful impact on the race. Indeed, the Ipsos/FiveThirtyEight poll found that almost no-one has changed their voting intention, even as a clear majority thought Biden won. But that is likely good news for the Democratic nominee, since, if the debate was an effective tie, that benefits the candidate who is currently in the lead.
The state of the race
There is no other way about it, Donald Trump is losing the presidential election. While the outcome remains uncertain and a late shift is always possible, time is quickly running out.
We have exactly a month left until election day and Biden leads Trump by 7.6% in the FiveThirtyEight national polling average. This time last month, he led by 7.5%, in August by 8.2% and in July by 9.6%. The race has tightened since mid-summer. It has not tightened anywhere close to fast enough.
Looking at the polls in key states, the trendline for the president happens to be even worse. The table below shows how the polling average has shifted in the thirteen battlegrounds I covered last month.
Data: FiveThirtyEight
In seven of the thirteen battleground states, Trump has lost ground over the past month. In most cases, the change has been very minor, yet Biden was already ahead and Trump is the one who needs a breakthrough. If the polls are exactly right (they won’t be), Biden wins 369 electoral votes, with 270 needed to secure the presidency. Now, let us assume a uniform polling error in Trump’s direction. If we take 5% off Biden’s margin in each state, he still wins 279 electoral votes. But polling errors can occur in either direction, so let’s say the polls are underestimating Biden by only 3%. In that event, he carries every battleground and wins 413 electoral votes, a historic blowout.
Covering the Trump/Biden horse race, I can’t help but feel a sense of deja vu. During the 2019 UK general election, the evidence throughout the entire campaign pointed to a comfortable Conservative majority. About a month before polling day, YouGov released its much-awaited MRP model. It predicted the Tories would win 359 seats.
Naturally, nervous progressives took comfort in YouGov’s stated margin of error. It’s final MRP, released mere days from election day, predicted 339 Conservative seats, yet with a plausible range of 311 to 367. To that effect, the most likely event was a small Conservative majority but the range of possibilities extended from a hung parliament to a Conservative landslide. In the end, Boris Johnson won 365 seats, just two shy of the upper limit in the model. For what it's worth, YouGov recently released its MRP for the US presidential race. It predicts Biden to win 350 electoral votes, with a Trump upset still possible in the margins.
So, the most likely outcome at this point in time is a sizable Biden victory. A narrow Trump win remains a possibility, but it is no more probable than a Biden landslide. Right now, Trump is ahead by less in South Carolina (a state he previously won by 14% and which I am not even including in my list of battlegrounds) than he is behind in Michigan or Wisconsin, among others.
Try as hard as you might, it is difficult to find good news for the Trump campaign. In the second edition of US Election Watch (back in mid-July), I argued that Biden will most likely win unless at least one of the following happens:
The economy rebounds far better than expected
The pandemic largely fades away
The Biden campaign implodes amid a scandal
So far, only the first has partially come to pass. The economy has improved, with unemployment falling and household spending rising. But it remains doubtful whether it is rebounding fast enough, or whether it will before election day, as economic recovery ultimately relies on successfully controlling covid. After all, no matter what the public guidelines are, people are less likely to travel and to participate in economic activity if they feel it might threaten their health.
On that front, over 200,000 Americans have died from the coronavirus. Meanwhile, both deaths and cases have been stagnant for months. And while Trump maintains a small lead on who is best to handle the economy, Biden is miles ahead when it comes to managing the pandemic.
What else is there? The president’s approval rating, at 44%, is only one point above his national polling. His disapproval rating is two points higher than Biden’s national support. Biden’s favourability, at 49.5%, is at an all-time high, so his campaign is hardly imploding.
The Trump campaign, meanwhile, is reportedly running out of money. Having spent big throughout the spring and early summer, to no discernible effect, it has now been forced to cut back advertising in a number of key states. At the same time, Democrats are pulling in record figures. Following the death of Ruth Bader Ginsburg, the party’s candidates raised more than double the annual GDP of Tuvalu, some $91 million, in 28 hours. Mike Bloomberg, the billionaire and former presidential candidate, recently committed to spending $100 million to defeat Trump in Florida, as the Biden campaign itself continues to outraise their opponent. In August, its fundraising advantage was more than two to one. It claims to have done even better in September.
There is one area where the Trump campaign is outperforming their opponent, the traditional ground game. While Democrats are only just resuming in-person canvassing, having previously suspended operations due to the pandemic, Republican volunteers have been knocking on doors for months. The result? A notable GOP advantage in voter registration - ‘Of the six states Trump won by less than 5 points in 2016, four - Arizona, Florida, North Carolina and Pennsylvania - permit voters to register by party. In all four states, voter registration trends are more robust for the GOP than four years ago’. There is a catch, which is that this will only ever make a difference in a close race. At the moment, Biden is too far ahead in too many places.
Just in case there is any doubt, Donald Trump can very much still win. I am probably not qualified enough to put a number on it, but there are statistical models built to do just that. As of this morning, Trump’s odds of being re-elected are 20% according to FiveThirtyEight, 17% in Decision Desk, and 13% in the Economist forecast.
People tend to be bad at comprehending probabilities between 1/10 and 9/10, but 10-20% is quite a lot more than zero. While one in five events are self evidently unlikely, they still happen all the time. If you are a distraught Republican, do remember that Trump’s odds are at least twice as good as getting a double pair in five-card poker. And if you happen to be an overconfident Democrat, realise that they are about as high as the chance of shooting yourself in Russian roulette.
In practical terms, the fact that Biden is polling at 50% or more in enough states to get 270 electoral votes means that Trump’s base mobilisation strategy is simply not good enough to win. The president has a high floor, in that he has never averaged below 40% nationally, but he has to start cutting into Biden’s support.
When does Trump have to start making a comeback? Yesterday would be a start. As I have been noting for the last few weeks, people are already voting. Over three million have cast their ballots, many of them in battleground states. In Wisconsin, those who have already voted make up 15% of 2016 turnout. If the race tightens in the final weeks, Biden’s early advantage could prove decisive. To win, Trump needs to quickly close the gap, or hope for a major polling error.
