A Political Scientist's Look at the 2020 US Elections

This will be a technical presentation on the election as of the information we have today but things are changing by the hour in the US presidential election.

I will divide important factors affecting the 2020 election into two types. First, there are factors that political scientists, historians and economists believe are well understood regularities. These are predictable relationships between underlying conditions or historical patterns and election outcomes. There are, however, some new things in the atmosphere this year that I would categorize as uncertain in their impact.

Historical patterns relating economic performance and incumbency to election outcomes are well understood and can be applied to 2020 without much difficulty. Also understood is the way that Donald Trump won four years ago--in somewhat unusual circumstances that put him in a precarious position in 2020. The second category of factors is more difficult to comprehend. One factor is the degree to which voters blame Trump for the spread of the pandemic in the US or view him as managing it well. Second is the very unusual economic situation in the US with the economy contracting and millions of jobs lost. That is obviously connected to the pandemic, and voters' views of the economy typically correlate with election outcomes, but the fact that this economic downturn is related to a pandemic for which Trump may not be seen as fully responsible complicates things. Finally, voters are also thinking about the protests about the treatment of black Americans that have broken out all around the country and Trump's handling of them.

Incumbent presidents have run for reelection 32 times in US history. 22 times the incumbent won and 10 times he lost, which is about a 70% success rate. The last loss was George Bush Sr., then Jimmy Carter before him in difficult economic times. Before that, 1976 was an unusual election where Ford had not been elected; he had taken office after Nixon resigned. Those are the three most recent cases, about a generation ago, and before that you have to go back to the Great Depression in 1932. They are exceptions to the rule. Not knowing anything about our current president or situation, we would think the president would be more likely to win.

However, other factors are pushing in different directions. One is how long the party in power has been in office when reelection is sought. After one term, the president was reelected every time from 1964 to the present except in 1980—six wins to one loss. When the president's party has been in office for two or more terms, the president has lost every time in the same time range except in 1988—one win to seven losses. We might expect Trump to be reelected based on this as well.

That is balanced by the way Trump won--which is to say in the Electoral College. The popular vote does not determine who wins an election in the U.S. Hillary Clinton won about 3 million more votes in the popular vote but Trump managed to flip six states that had gone to Obama in 2012, but in some cases narrowly. The three most important states are Wisconsin, Michigan and Pennsylvania. Trump won these states by less than one percentage point, by a combined total of 80,000 votes.

As an incumbent seeking reelection, this is a very difficult position for him because he did not win the popular vote and his Electoral College victory was dependent on these very narrow victories in three states and no gains elsewhere. Iowa, Ohio and Florida are also swing states that flipped to Trump in 2016 and Trump is very much in jeopardy in all three this year. He begins without an overwhelming victory that he could build on in 2020 and to be successful, his campaign would probably have to replicate what happened last time. He will need to win all of the states he won last time. That's a difficult task given the situation. The nature of his victory tempers the pattern of incumbents being a cinch for reelection.

Because of the surprising nature of his victory, he did not begin his first days in office with high approval ratings. The public was immediately negative about his performance and that has remained remarkably stable. It is surprising that his approval ratings have hardly moved. He has a "low ceiling" but a "high floor" meaning that his support have kept this minimum level stable. His approval ratings rarely drop below 40% and have rarely moved above 43-44%. When the pandemic began in the US in late March and early April, there was a slight uptick in his rating. As the reality of the pandemic and other issues crept in, his rating fell through the summer. In the last couple of months, his rating has recovered a little but he remains in a difficult position. Even a change of 1-3% could change the election.

There is a strong relationship between presidential approval in June and the share of the vote won by the incumbent party in presidential elections. When the President has higher approval ratings, his party tends to win by large amounts. Trump's approval was about 43-44% in June. In the scatter plot, that lines up with a vote share of around 50%. This is the tipping point where a president can be just popular enough to win or just unpopular enough to lose. It seems very unlikely that Trump will win the popular vote but it is possible that his rating will be just high enough and that other factors will come together to give him a chance to win the electoral college.

A recent national survey showed 41% overall approval and 58% disapproval of Trump. He received more approval than disapproval for his handling of the economy, while it is lower for court appointments, the coronavirus and the Black Lives Matter protests. His handling of the last two received very low levels of approval: 38% and 37%, respectively. His strength is his handling of the economy and this has been a good predictor for election outcomes. This could make him competitive against Joe Biden.

What is the relationship between the economy and presidential election outcomes, and is Trump in this territory that would make us think he has an opportunity to win? Even small positive change in the GDP from the first to third quarter of an election year tends to result in presidents winning reelection. When growth is not as strong, or when it is flat or negative, presidents tend to pay the price and are not reelected. Where does Trump stand in terms of this? Third quarter results are not in yet, but forecasts from financial firms and think tanks are suggesting that the GDP contracted anywhere from -1.5 to maybe -4 or -5 points. That puts Trump in virtually unprecedented territory, making it very hard for him to win reelection. However, many Republicans and some independent voters don't blame Trump for this, due to the coronavirus pandemic, which they blame on other factors or parties. However, he is in a very difficult position to win the popular vote and is still in a difficult position to win in the electoral college given the historical relationship between the economy and reelection.

National surveys have consistently shown Biden ahead by anywhere from 4% to 10% for over a year. Today, it stands at a gap of about 9 points. That is a very large gap, especially for an incumbent president. If that holds until election day, it would be the worst popular vote defeat for an incumbent president since at least 1980.

The polls also showed that Hillary Clinton was going to beat Trump four years ago and they were wrong, but I think the situation is different now. The trend lines show different patterns between 2016 and 2020. There was greater volatility in 2016, And at certain points, Clinton and Trump were tied. Trump was also in fact ahead at one time, so it was less clear that Clinton was in the lead. She only had a lead of 3 points, which is almost exactly where she finished in the national popular vote. In contrast, Biden is ahead by 8 or 9 points and this may have grown since Trump contracted COVID and the first presidential debate. The larger and more stable gap in 2020 makes me confident that Trump is in real trouble. In addition, Trump is the incumbent, he is well-known, and the election is largely a referendum on his performance and the public has very strong views on him. The polls seem to more clearly show the situation for this election. Swing state polling also shows that Trump is behind.

Because of the electoral college, the battleground states can be a tipping point between Democrats and Republicans and they're getting a lot of attention from the candidates. Trump is trailing in most polls in most of these states. It appears Biden has more than the number of electoral college votes he will need but there is still time before the election.

Another factor to keep in mind is that the electoral college is disadvantageous to Democrats. Small, less populous states are over-represented while more populous states are less represented. That means the Democrats need to win more than just a majority of the popular vote to win the electoral college. When candidates are tied in the popular vote, according to The Economist's simulation, the Democrats generally do not have enough electoral college votes. It seems they need to get to about 52% to start winning and to about 54% to win in the electoral college on a consistent basis. It's not a 50-50 race but 53-47 or 54-46.

There are a few other things happening in this election that introduce uncertainty. How will the pandemic affect voter turnout? Many Americans have decided to vote by mail, something used only by about a quarter of voters in 2016. Voting is already taking place. About four million votes have been cast nationwide already. We don't know whether this will lead to higher or lower turnout. Higher turnout generally provides advantages the Democrats. One clue is the midterm election turnout. These are elections for Congress and state offices. In the last one in 2018, turnout was about 50%: the highest turnout in a midterm election since 1912. The 2018 midterm election suggests a lot of energy and engagement by voters. We do not know what turnout will be but some predictions see 150 million people voting, a turnout rate of about 62% to 63%, near a record-high for the last century or so, which would presumably help Joe Biden. The early voting by mail has been very strong.

The entire House of Representatives and a little more than one-third of the Senate are also on the ballot this year. Democrats have about a 40-seat majority in the House, meaning that Republicans would need to gain about 20 seats to take back the House. That seems very unlikely in this election environment. Republicans are unlikely to make big gains when the president is unlikely to win the popular vote. Democrats are in a pretty safe position to retain their majority in the House. In the Senate, Republicans have 53 seats and Democrats 45 or 47 if you include the independent senators. Biden winning would flip the Senate to the Democrats. The prospects are quite good for them to take the Presidency, the House and the Senate. Many of the Senate seats that are up this year are being defended by Republican Senators, putting them at risk. The Democrat seats that are up are mostly in blue (Democratic) states, so they're not in trouble. There is only one state where the Democrat is likely to lose--Alabama. There are plausibly 5 to 7 states where a Republican Senator could lose to a Democrat, giving the Democrats the majority in the Senate even if they lose the Alabama seat.

The most likely outcome is Biden wins the popular vote by a significant margin, giving him a good chance in the electoral college to become President, Democrats hold the House by roughly the same margin, and in the Senate, the Democrats have an opportunity to win a majority. The majority may not be enough to enable them to effectively govern because the Senate filibuster rule typically requires a super-majority of 60 votes for most legislation, but it would allow a Democratic Senate to do things like nominate and approve federal judges. The Senate results may also be delayed for some time after the election.

ONO Yoshikuni: I have some questions. The first is about voter mobilization. During the Obama campaign, young people and black voters were very active. To some extent, the Clinton campaign did not fully mobilize those voters. I am curious about how Biden is doing, especially among young voters. For example, they cannot campaign on college campuses due to the pandemic, so how it this aspect proceeding?

Barry BURDEN: This is a good question. The Democrats have decided to stop all face-to-face campaigning and have relied more on digital advertising. Biden has raised plenty of money and he has a lot of volunteers texting and making phone calls. Conveniently, young voters the Democrats want to reach are more reachable on digital technologies. Young Democrats are telling us in surveys that they are hearing from the Democratic campaign, mostly through digital routes.

Republicans are using a more traditional mix of mail and face-to-face campaigning. Trump was visiting many states until he caught COVID and his people are visiting people's homes. The Democrats want to make sure they don't make the same mistakes the Clinton campaign made four years ago. They didn't play a strong ground game, especially with regard to black voters. I think there is more genuine enthusiasm for Biden because he was Barrack Obama's Vice President and that connection creates warmer feelings for black voters. Young voters dislike Trump. They are eager to vote despite not being so excited about Biden. In a survey, Trump supporters say they are voting for him because he is their preferred candidate. Biden's voters tell us they are not great fans of Biden; they only want to get rid of Trump. That has brought the Democratic Party together.

ONO: I guess many voters will not want to go to polling stations so many will mail their ballots. I am curious about the impact of the mail-in ballots. There is debate about their validity.

BURDEN: There is a real risk in votes being returned by mail. Between 1% and 2% of mail-in ballots are not counted because something goes wrong. This year, many people will be voting by mail for the first time and new voters make mistakes that can cause their ballots to be rejected. The Democrats are taking a risk in putting so much effort into mail-in voting. But there is a partisan divide this year, as you hinted, where Democrats want to vote by mail and their leaders are telling them to do so. Republicans are more mixed. Trump has disparaged voting by mail and has been encouraging his voters to vote in person. This is new: before the pandemic, Democrats and Republicans basically voted by mail equally.

ONO: I am interested in the effect of Kamala Harris as a female candidate for VP on the election results.

BURDEN: She's not only female but also of mixed ancestry. That gives her a different appeal than other female candidates. There have been other female candidates but none of Harris' predecessors were successful. Vice Presidents are getting more attention this year as both Presidential candidates are in their 70s. I don't think most of the public knows much about Kamala Harris. I don't know what her impact will be. If voters are reluctant to see a woman in the White House, they may be reluctant to vote for Biden as she could become President. Biden has hinted that he may not run for a second term, and she would then be the heir apparent.

WATANABE Tetsuya (RIETI Vice President): If we focus on the six battleground states, what will be the main deciding factors in voters' behavior, allowing voters who voted for Trump to flip?

BURDEN: In the three key swing states of Wisconsin, Pennsylvania and Michigan that Trump won so narrowly in 2016, there are a few key groups to look at. One group voted for third-party or independent candidates in 2016; these candidates got about 6% of the vote. Trump won these three states and others, but without winning 50% of the vote. This year, he will need 50% because the third-party candidates will have a much smaller presence. Where will those third-party voters go this year? In our research, the subjects that were surveyed are saying they are likely to vote for Biden by a margin of almost two to one. Because they comprise 6% of the vote, that's a very large share going to Biden. It puts him over the top in some of these key states.

Another key group that Professor ONO mentioned are African-American voters, who are very important especially in larger urban areas like the cities of Milwaukee, Detroit and Philadelphia. Higher black voter turnout will help the Democrats. The Democrats seem to have focused on them this time, in comparison to 2016. Their turnout level remains uncertain.

A third group to watch are voters in rural communities, who supported Trump very strongly in 2016, especially those with lower levels of education. They turned out at much higher rates than in the past. They look like they are not going to be as supportive this time around. Just enough seem to be moving back to the Democrats to cost Trump the election. These are all small factors of a few percentage points here and there, but because of the narrow margin of victory in the swing states, he cannot afford these various losses.

WATANABE: What is the main reason why rural voters are now more reluctant to vote for Trump? Is that economic policy performance or the pandemic or other factors?

BURDEN: I think much of it is frustration that led them to support Trump which has not been satisfied over the last four years. The things that attracted these voters to Trump four years ago were his message about shaking up Washington and "draining the swamp." Many of them felt that Washington was out of touch with their needs and did not respect them, so they liked that language. Trump has not fully delivered on that in their view. He is now a part of Washington and is not the outsider candidate he was four years ago. Rural voters are also frustrated by the cost of healthcare and healthcare access. Obamacare was not working for rural voters. Rural healthcare providers were closing down and the cost was going up every year under Obamacare. Trump promised to dismantle the Affordable Care Act and build a better healthcare system, which he has not done. Rural voters are not happy with Obamacare or with what Trump has done. There is also the mixed effect of Trump's unresolved trade wars, especially in farming communities dependent on international trade. I think Trump's troubles with rural voters started before the pandemic and racial protests.

YANO Makoto (RIETI Chairman): Many outside factors or information control factors exist, including the possibility of Russian intervention. I think James Comey's treatment of the meddling and Clinton's emails were both critical to the 2016 election result. In this election, informational interventions, and Trump's treatment of QAnon, the Proud Boys and Antifa and such groups are all external factors that may have a sudden, last-minute effect on voter behavior. There are all sorts of manipulations that could play a role in determining the election. What is your assessment of these factors in this election?

BURDEN: There already are some disinformation campaigns this year. Some are coming from the Russian government or are sponsored by Russian actors. It looks somewhat different than four years ago but it is happening on social media and elsewhere. The Iranian and Chinese governments are also sowing disinformation, not necessarily to hack the election but to create discord and confusion and lack of confidence. I don't think the social media companies are doing a good job of containing that. I think this year will be different because the Russian actors were more supportive of Trump than Clinton in 2016 and the public likes Biden more than they liked Hillary. He does not have the same baggage or scandals. Where Trump has tried to tag him with things like the stories about Hunter Biden, it hasn't worked. I don't think the disinformation will be as effective this time around, and that is disadvantageous to Trump.

YANO: I would like to know your views on whether Americans in general and the American media and government are more aware and immune to the kind of information intervention we saw in 2016.

BURDEN: Americans on Facebook who saw posts in 2016 did not realize they were coming from Russian actors and the public has since learned about that. The Senate Intelligence Committee has done a series of bipartisan reports that have been very good on Facebook ads that were deceptive and probably illegal and paid for by Russians in rubles. The intelligence community also did some reporting. The good thing is that election officials in the US have spent the past four years trying to harden the election system. Despite Trump's personal dismissal of outside meddling, the Department of Homeland Security has helped the states implement significant protections. I feel much more comfortable that the election system has the right protections in place. It's very hard to say what is going to happen, but there are likely to be some different elements this time around.

The U.S. is a big, diverse, pluralistic country so there are a lot of subpopulations that can be persuaded by targeted messages. The nationwide election system is actually 50 state elections with internal counties etc., which provides a lot of points of entry for a bad actor to reach a group or burrow into a state and try to have some influence. Even with all these protections in place, it is a permeable system, perhaps different than the systems many other countries have.

ONO: Many of the efforts to discredit Trump are coming from the Republican side, which is surprising. Why has this happened?

BURDEN: It's very unusual. Former officials from previous Republican administrations and several former chairs of the Republican National Committee have endorsed Biden. I think it's an indicator of how much the Republican Party has changed in four years. It's really Trump's party rather than a traditional Republican Party. If Trump wins it will not be because of help from the Republican Party; it will be because of him and his most hardcore supporters. During his election campaign four years ago, he was not cooperating with the Republican Party. He viewed both the Democratic and Republican parties as his opponents. It has become part of his brand and that resonates with many of his supporters. But without the party coming together, it is hard to win an election. Right now, surveys show that people who consider themselves Republicans are very supportive of Trump but their turnout levels may be lower.

The Senate will vote on Amy Coney Barrett's nomination and the question is whether this is going to help Republican Senators who are up for re-election. I think this depends on the state, as does Trump's effect. In a state like Georgia, which is a swing state this year, one of the Republicans is running ads showing how similar she is to Trump. If a state leans red as Georgia does, this may help, as appears to be the case in Iowa, where Trump won by a large margin in 2016.