Why Trump's Reelection Odds are Lower than You Think

Will Trump still reside at 1600 Pennsylvania Avenue in 4 months? I have my doubts. Evan Vucci / Associated Press

On Tuesday, which marked six weeks till this year's federal elections, I asked on Facebook what people thought Trump's chances of reelection were. The feedback was not overwhelming (Only four people answered, including a text exchange) but the most common response was that Trump is a moderate underdog with the Supreme Court opening & the potential for poor debate performances by Biden leaving the door open for this to turn into a toss-up. The one dissenting opinion had Trump as a solid favorite, citing incumbent advantage (four of the last five have won), a strong pre-COVID economy, & lack of excitement around Biden. 

All of the answers had some solid reasoning behind them, but I wasn't surprised to find that they all gave Trump better odds than I do (I'd put him at around 25%). At the risk of being proven a fool in six weeks, here are my thoughts on why it's unlikely:*

*A quick note: References to current polling/odds are often linked to FiveThirtyEight, or The Economist, whose polling & election models are constantly updating, thus numbers may have changed slightly since I posted this. 

• His approval has been underwater his entire presidency, even in the pre-COVID economy. On January 1st of this year, before most Americans were even aware of COVID-19, Americans disapproved of Trump 53-43. What are those numbers now? 53-43. In fact the only time his disapproval rating has dropped below 50% since March of 2017 was after he declared a national emergency in response to COVID. Americans may approve of his handling of the economy but that hasn't stopped a clear majority from thinking Trump has been bad at the job throughout his presidency.   

• Remember the four of five incumbents who won reelection? At this point in their first terms they all had approval ratings of at least 49%, Trump has never hit that number. They all had net approval ratings of at least +2, as mentioned above, Trump is at -10. Now to be sure, the last two presidents to lose an election had worse numbers, so I'm not saying Trump is completely done for, but the idea incumbency is some great advantage for him doesn't pass the smell test.

• The things that people keep saying will help Trump haven't. Protests/riots? Biden's lead peaked in in June & July at the height of the unrest, not surprising considering polls showed voters trusted Biden more on race issues & police reform. What about the Supreme Court opening? It's too early for hard data but we had elections shortly after the Kavanaugh confirmation & Democrats gained forty-one seats in the House. Yes, Republicans picked up a net of two Senate seats but three of their four wins of previously Democratic seats were in beet red Indiana, Missouri, & North Dakota. The races we should look at are the ones in purple states. Democratic incumbents won in key swing states like Michigan, Pennsylvania, & Wisconsin, & took Republican seats in Arizona & Nevada. Their only notable loss was Florida. So there's no evidence that a Supreme Court nomination helped Republicans then, & that was with them replacing a conservative Justice. Is replacing a liberal icon with someone who will push the Court much further to the right really going to be helpful when 92% of Republicans are already on board with Trump, a third of 2016 Trump voters were either pro-choice or had mixed feelings on abortion, & Democrats are more motivated by the Supreme Court? Color me highly skeptical. 

• Likewise the debates seem unlikely to change the race in any significant way. Debates only tend to matter at the margins, potentially deciding close races as the first televised debate between Kennedy & Nixon supposedly did in 1960, but not causing massive changes to polling numbers. Biden isn't the strongest debater, but that actually might help him here. Hillary Clinton was widely considered a much better debater than Trump, so when that proved true in 2016 she merely met expectations. This time both candidates will have fairly similar, if similarly low, expectations put on them. Further Biden is actually better one-on-one than in the crowded debates that marked most of the Democratic primary. He won the Vice Presidential debates in 2008, & 2012, & in his lone one-on-one debate with Bernie Sanders he completely held his own. It makes sense that a man with a life long stutter would be more comfortable on a stage with a single opponent, knowing exactly when he will have to speak, rather than surrounded by a half dozen competitors who might attack at any moment. Biden is less likely to fall apart in the debates than many people think, & even if he does it will only shrink his lead rather than erase it.

• Changing demographics. Liberals put too much faith in increasing diversity delivering election results in 2016, & they shouldn't be viewed as a guarantee now, but it's worth noting that if every group votes at the same rate, & for the same party, as in 2016 Biden would actually win because white people now make up a smaller share of the electorate. Of course numbers won't be the same, but these changes are definitely to Biden's advantage.  

• The race has been remarkably stable. Biden has held a 5+ lead for five months straight. Even before COVID Biden had a 4 point lead. The day George Floyd was murdered Biden led by 6. He's now up 7. Biden has a sizeable national lead, & time is running out for Trump to change that. 

• You might have read that last bullet point & thought, "Who cares about the national race, Clinton (& Gore) both won the popular vote but lost the Electoral College. That's all that matters," & that would be a fair point. A bit harder to win when losing by 7 nationally, or even 5, than 2 (2016), or 0.5 (2000). Trump still has an Electoral College advantage, the most likely tipping point state (the state that decides the election) is Pennsylvania where Biden currently holds a 4.6% lead, but a 3 point advantage in the EC doesn't do much good if you're down 7. He's going to need to cut Biden's lead roughly in half to have a chance, & he only has thirty-nine days to do so. 

• "The polls were wrong last time though," someone is surely shouting. Yes, the polls were notably wrong in a half dozen of decisive states (National polling was actually more accurate than in 2012), & they won't be perfect this time either, but a number of pollsters have fixed the biggest reasons for those misses - a lack of educational weighting. Polling involves weighting, or adjusting, results to more accurately reflect the populace being polled. In 2016 only "20 percent of battleground state pollsters" were weighting by education, which was a problem since college educated people are more likely to respond to phone surveys & the Midwestern states that decided the election have an especially high percentage of non-college educated white voters. Thus most of those polls whiffed on the result. This time nearly half of state pollsters are weighting by education, that's still too low, but it is a significant improvement. 

• The other big issue in polls missing in 2016 was the huge number of undecided voters. At this point four years ago about 20% of voters were either undecided or leaning toward third party candidates, & Trump went on to win a clear majority of the 13% of voters who decided in the last week before the election. If the numbers had been even Hillary Clinton would be president right now. Polling now shows undecided voters ranging from 5 to 10%, leaving much less uncertainty about this race. 

With all that said, it hardly means the election is over. FiveThirtyEight had the election model that gave Trump the best chance of winning in 2016 (29%), & they currently have his odds at 23%. That's not a huge difference but beyond my belief that polling should be more accurate FiveThirtyEight has changed their model to allow for more uncertainty & that's currently boosting Trump's odds. For comparison, The Economist, which didn't run a model in 2016, gives Trump just a 15% chance of winning reelection. Normally I might make my best guess by aggregating the two, I think they both have solid reasoning behind their approaches, but the added question about how many mail-in ballots will be disqualified has me taking a slightly more conservative approach. Plus even The Economist's odds are roughly the same as predicting the number you'll roll on a six-sided die, not likely, but far from impossible. 

Still, while Trump should by no means be ruled out, I think it's clear he's a significant underdog, & unless there's an unforeseen gamechanger on the horizon Joe Biden will likely be president in four months. 

If I'm right, we'll probably see "Diamond" Joe Biden in the Oval Office next year. The Onion