The 2020 data says, ‘that was close’
Joe Biden was elected the 46th president of the United States. This will be official once the electors cast their votes on Dec. 14. With enough of a cushion, faithless electors will not be able to change the outcome.
In spite of polls indicating that the election would be a rout, with President Trump having little chance to win many, if any, of the key battleground states, he secured convincing victories in Ohio, North Carolina, and Florida. In fact, he is the first sitting president since 1892 to win Ohio but not win reelection.
A deep dive into the data indicates that when comparing the 2016 and 2020 elections, Trump’s 2020 loss was by a much narrower margin than Hillary Clinton’s 2016 loss.
In 2016, Clinton lost three critical states (Michigan, Pennsylvania and Wisconsin) by under 78,000 votes out of just under 14 million votes cast. If she had succeeded in winning all three of these states, she would have narrowly won the election.
In 2020, Trump is on track to lose three critical states (Georgia, Pennsylvania and Arizona) by a combined 76,000 votes out of just over 15 million (and still being counted) votes cast. He also won North Carolina by around the same number of votes, a hotly contested battleground state, nearly matching his total losing margin for Georgia, Pennsylvania and Arizona. If he had succeeded in winning all three of these states, he would have won the election, in spite of losing the popular vote (as he did in 2016).
These 76,000 votes would have required half of these people switching their vote from Biden to Trump, representing just one-quarter of one percent of the voters (or one out of every 400) in these states. If this small number of people had a last minute change of heart, or even accidentally darkened the wrong circle when casting their ballot, the results would have been different with a Trump reelection.
Biden also won four states by less than one percent (Pennsylvania, Georgia, Wisconsin and Arizona), while Trump’s closest winning state will be North Carolina, which should be over one percent. Biden took a page out of Trump’s 2016 playbook to win key states by just enough, but no more.
The votes cast for neither the Republican nor the Democrat candidates changed dramatically between 2016 and 2020. In 2016, the three critical states had over 5 percent of the total votes cast for third party or write-in candidates. In 2020, the three critical states had just over 1.3 percent of the total votes cast for third party candidates or write-in candidates, a drop of over 70 percent. This indicates how serious voters were about casting their ballot for either Biden or Trump, and making their votes be heard.
It is easy to play Monday morning quarterback and say what could have been. Without COVID-19 and the ensuing economic meltdown in the spring, Trump would have likely enjoyed a landslide victory, certainly winning the three key states he lost listed here, and perhaps a few more. Alan Lichtman’s two economic keys would have turned favorable for Trump, forecasting a Trump reelection. Indeed, even with the backdrop of a global pandemic, resulting in over 240,000 COVID-19 deaths and over 10 million cases to date, he was literally a few thousand votes away from snatching victory from the jaws of defeat.
Pollsters once again underestimated support for Trump, forecasting a Biden rout, not foreseeing that his victory would take four days after Election Day to be called by the Associated Press and other networks. Are pollsters losing their ability to accurately capture voter preferences from samples, or did Trump’s style make it difficult to accurately capture how people would cast their ballot? We will know more about this in 2024.
Whatever happens over the next 10 weeks until Inauguration Day, the final voting tally will indicate that the result, though widely forecasted, was far closer than many believed that it should have been. Indeed, in spite of what is said, President-elect Biden’s victory was much narrower than Trump’s 2016 victory. He was that close to winning a second term.
Sheldon H. Jacobson, PhD, is a professor of computer science at the University of Illinois at Urbana-Champaign. He applies his expertise in data-driven risk-based assessment to evaluate and inform public policy. He is the founder of Election Analytics at the University of Illinois, a STEM learning laboratory for election forecasting.