Joe Biden has very little chance of losing the national popular vote. No candidate with a double digit lead at the end of July has ever lost the national vote in November since Gallup started polling in 1932. The unpredictability of the pandemic and the associated economic downturn creates less certainty about the tally this year. But barring the “disappearance” of the virus predicted by President Trump or some galvanizing act of leadership by the incumbent, the Democratic nominee will likely see his current polling lead shrink only a little as he crosses the finish line with a 7% to 8% margin. If turnout is at the high end of the range between 2008 and 2016 (around 60%) that margin will translate to winning by 10 million votes.
Under this scenario Donald Trump still has a one in five chance of winning the electoral college and the Presidency.
Using the average of recent state polls, we simulated the 2020 election many times, taking these polls as a starting point and adding in historical variation in turnout, third party support level, and other factors that could cause actual voter results to differ from the polls. As things stand in each state, Trump has roughly a 20% chance of winning the Electoral College while losing the national popular vote margin by an average of +8%, which translates to 10.3 million more votes.
The reason Trump still stands a puncher’s chance is that in the last 20 years partisan division has been defined starkly in racial, education, and gender terms. As a group white males without a college degree have voted for the Republican candidate in every presidential election since 1968. Their solid support for that party’s nominee gives Donald Trump a puncher’s chance of winning the electoral college this November even if he loses the national popular vote.
White males without a college degree currently represent around 25% percent of the electorate, but there are large differences across states. In the 29 states where white males without a college degree account for at least 30% of the population, since 2000, the Democratic nominee has won only 14 of them in at least one election and won only 9 in 2016. Across recent elections, the Democratic candidate has only a 39% win percentage in these states compared to a 53% win percentage in states where the fraction of white men without a college degree is below 30%.
Under current polls, Trump is likely to win around 192 electoral votes, well short of the necessary total. However, in simulations from our model where he does amass the necessary electoral votes, Trump is able to just win by holding the 12 states where he leads under current polling by at least 10 percentage points, keeping the 12 states that he won safely in 2016 and that historically lean Republican where he trails or holds slimmer leads today (Alaska, Texas, Utah, Montana, Arkansas, Georgia, Mississippi, Missouri, South Carolina, Ohio, Iowa, and Tennessee), and eeking out enough electoral votes in the current battleground states (Arizona, Pennsylvania, Florida, Wisconsin, and North Carolina).
Ten of these 17 states are those where white men without a college degree represent 30% or more of the population: Arizona, Arkansas, Iowa, Missouri, Montana, Ohio, Pennsylvania, Wisconsin, Texas, and Utah. In Florida, another of the 17 states, 29% of the population is white men without a college degree. The 6 other states, except for North Carolina, are traditional Republican strongholds (Alaska, Georgia, Mississippi, South Carolina, and Tennessee).
The electoral landscape for the Trump strategy is then focused on a two step process of first shoring up the Republican base and then capturing the battleground states, especially in relation to motivating his base of white men without a college degree. He will hope that Biden’s national victory comes overwhelmingly from college educated voters, especially women and nonwhites, running up the vote totals in safe and populous Democratic states like California and New York.
Winning the 12 states he currently leads by at least 10 percentage points would give Trump 72 electoral votes and holding onto the 12 states he won safely in 2016 would provide another 132 electoral votes. He therefore needs at least 66 electoral votes from the battleground states. This makes Florida and Pennsylvania both essential because if Trump doesn’t win these, he won’t amass the necessary votes to cross the 270 threshold. If he can win those two states though, Trump can take the election by winning any two of the remaining states among Arizona, Wisconsin, and North Carolina.
Trump’s revealed ad strategy perfectly reflects this process. Currently, he has ramped up spending in states that he won in 2016 by five or more percentage points and where his base is a large portion of the electorate such as Iowa, Ohio, and Georgia. Trump has also planned for at least $100 million in television ads in the key battleground states of Florida, North Carolina, Arizona, Wisconsin, and Pennsylvania starting labor day and lasting till election day.
If Biden’s 10 million votes come from a greater percentage of victory among women than Clinton’s votes (she won females by 12%) and among nonwhites (she won them by 58%), then correspondingly Trump may increase his white male without a college degree share from the 70% he obtained in 2016 to a higher number.
If Trump wins this group by 72% while losing the female and nonwhite votes by larger margins than he lost them to Clinton, he can still win the electoral college following the path described above. Specifically with this strategy he can amass enough votes to win by holding his base and capturing enough of the battleground states. These 17 states happen to represent only 33 percent of eligible voters but 40% percent of all electoral votes; the 5 battleground states represent only 10% percent of eligible voters but 16% percent of all electoral votes. These facts are a result of the disproportional way electors are allocated against the voting population. Therefore, the Republican base that is focused on the less educated white male population maps against a rational electoral strategy.
In this scenario Joe Biden could win the popular vote by 7% and 10 million votes and roughly equal Barack Obama’s 2008 margin of 7.2% and 7 million votes but still lose. Obama won the electoral college by 365 to 173. Here Biden would lose 259 to 279.
His margin, like Hillary Clinton’s national vote victory, would be tilted heavily towards the Democratic strongholds where white males without a college degree compose a smaller share than the national average percentage of the population. Scenarios like these give Trump a 20% chance of winning the electoral college even while losing by national popular vote by a large margin.
The Trump election strategy then depends on turnout by white males without a college degree in the 17 states where he currently holds a small lead or trails by a small enough margin to make victory possible.
Biden has to like his odds. But the statistics explain why Donald Trump uses every conceivable way to communicate to less educated white males that he is their voice. They also reveal that the electoral college inherently invites dividing the nation on gender, race, education, and geographic grounds.