How did Donald Trump win the 2016 Republican nomination and the presidential election? Three academics look rigorously at the data to explain how and why it happened.
3 August 2020
I have followed American politics closely almost all my life.
Aged 10, I asked my mother to wake me at 06:00 rather than my regular 08:00 so that I could know the result of the Kennedy / Nixon election as early as possible. I have stayed up all night for almost all subsequent presidential elections.
I was deeply dismayed in 2016 when Hillary Clinton lost to Donald Trump. However, unlike most people, I was not shocked.
I had read Nate Silver’s final forecast issued on the morning of election day. According to him, Clinton had a 71% chance of winning based on polls only, or a 72% chance based on his poll-plus model. This was a much lower probability for Clinton than almost all other forecasters. 28% probability outcomes happen regularly. That is why I was not shocked, but I knew it was an awful result for America and the world.
Since then, I have read many explanations of why Trump won. However, most of them are unsatisfactory because they come from the explainer’s “gut feel” and often seek to validate the explainer’s own political views.
This book is different because it is almost entirely grounded in data.
No author comes to any subject without personal views. However a good academic learns how to put these to one side when seeking knowledge. In my opinion, these authors have succeeded in doing that, mainly because they are academics rather than journalists or politicians.
The dust jacket has the following brief biographies.
John Sides is Professor of political science at George Washington University.
Michael Tesler is associate professor of political science at the University of California, Irvine. He is the author of “Post-Racial or Most-Racial?: Race and Politics in the Obama Era.”
Lynn Vavreck is the Marvin Hoffenberg Professor of American Politics and Public Policy at the University of California, Los Angeles. She is the author, with John Sides, of “The Gamble: Choice and Chance in the 2012 Presidential Election.”
The book comprises 246 pages (including the appendices to some of the chapters) + 87 pages of notes and index.
It has the following chapters:
The book has so many figures (graphs or charts) and tables that three pages are required simply to list them. This illustrates how much it is based upon data.
The book opens with a Trump rally on 9 March, 2016 in Fayetteville, North Carolina. A black anti-Trump protester, Rakeem Jones, was being led out of the rally by police officers, when he was punched by 78-year-old white American John McGraw who said of Jones “We don’t know who is, but we know he’s not acting like an American. The next time we see him, we might have to kill him.”
The authors mention several other violent incidents at Trump rallies, and the way that he encouraged attacks on protesters. The authors write:
McGraw’s comment “we know he’s not acting like an American” distills what the election was fundamentally about: a debate about not only what would, as Trump put it, “make America great again,” but who is America – and American – in the first place. It was a debate about whether the president himself, Barack Obama, was an American. It was a debate about how many immigrants to admit to the country. It was a debate about how much of a threat was posed by Muslims living in or traveling to the United States. It was a debate about whether innocent blacks were being systematically victimized by police forces. It was a debate about whether white Americans were being unfairly left behind in an increasingly diverse country.
What these issues shared was the centrality of identity. How people felt about these issues depended on which groups they identified with and how they felt about other groups.”
The rest of this chapter expands on this theme, at this stage without the data that the authors use in the rest of the book to understand what was happening.
The authors start by looking at the performance of the American economy and how this normally ties in with presidential approval ratings. For most presidents since Truman, the more positive the index of consumer sentiment, the higher the president’s approval rating.
For some presidents, the relationship is quite strong while for others the relationship is much flatter although still positive. Obama is the only president for whom the relationship is slightly negative.
Based upon historical relationships between the economy and voting patterns, looking at the data around the end of 2015 the authors would predict the election as a “toss up.”
The authors then cover the high level of partisanship in American politics, showing that voters' assessment of the state of the economy is heavily influenced by their party allegiance.
“YouGov / Economist polls conducted from June to December 2015 found that, among Democrats, 27% said they were better off financially than a year ago, 48% said that their finances were about the same, and 20% said they were worse off financially. By contrast, only 11% of Republicans said they were better off financially, while 43% said they were worse off.”
The authors point out that American politics are increasingly divided along race.
“The “racialization” of partisanship was underway even before Obama became a national figure. Americans’ partisan attachments became more closely aligned with racial attitudes in the post-civil rights era as politicians from the two parties increasingly diverged in both their policies and their rhetoric about race. But eight years of an African-American president accelerated and intensified racialization. This meant that the outcome of the 2016 election would depend not only on election-year fundamentals like the economy but also on how successfully the candidates could navigate these racial dynamics and mobilize a winning coalition.
The first major change in the party coalitions was the increasing Democratic advantage among nonwhites (figure 2.6). This was not preordained: for many years, many nonwhites – especially Latinos and Asians – had not consistently aligned with one political party. But that changed. Although there was no secular trend in Asian American partisanship from 2007 to 2016, the longer-term trend was clear: in exit polls, Asian Americans’ support for Democratic presidential candidates increased from 31% in 1992 to 73% in 2012. Latinos also came to identify more with the Democratic Party. Among Latinos, Democrats outnumbered Republicans by 23 points in 2002 but 36 points in 2016.
African-Americans’ identification with the Democratic Party strengthened as well, even though blacks had long been Democratic.”
The authors also present other data showing how attitudes to race increasingly separated Democrats and Republicans.
This chapter is entirely about the Republican Party. It looks at the internal divisions within the party, especially after the 2012 presidential election whose result shocked many Republicans even though the polling had predicted a victory for Obama.
“In a Gallup poll conducted 10 days before the election, almost 3/4 of Republicans expected Romney to win. Romney himself had written a victory speech but no concession speech.”
There is a very informative figure looking at the pattern of endorsements from Republican office holders (congressmen, senators and governors) for Republican presidential candidates from 1980 onwards.
This shows how in 2016 there were far fewer endorsements in the year before the election year than most previous campaigns. 2016 also had the lowest number of endorsements won by any Republican presidential candidate in the period up to the Iowa caucuses.
“The upshot of the 2016 invisible primary [the quest by candidates for endorsements from party leaders] was that party insiders could not identify one single candidate who stood above the others on both criteria: satisfactory on the issues and electable in November.”
Later chapters of the book show how this indecision allowed Trump to win the nomination. What happened to the Republican Party in 2015 — 2016 clearly lies behind the way that the Democratic Party coalesced rapidly around former Vice President Joe Biden after he won the South Carolina primary in February 2020.
This chapter looks at the share of news coverage received by Trump during the campaign for the Republican nomination from May 2015 onwards. It compares that with the amount of coverage received by the other major candidates, Ben Carson, Jeb Bush, Marco Rubio, and Ted Cruz.
The authors also look at the argument that much of the news coverage that Trump received was critical and negative. Their close look at the data shows that this view is incorrect. Much of the coverage was neutral or even favourable.
The authors also point out that, in the same way that Republican leaders failed to coalesce around any candidate, the candidates running against Trump for the Republican nomination failed to attack him.
In figure 4.5, the authors look at data from December to April 2016 for the volume of Republican attack advertisements, where an attack advertisement is any that mentions another candidate by name in a critical fashion.
As long as there were many candidates, most of the attacks were not attacks on Trump.
“The same pattern [referring to the previous paragraph where few of Trump’s primary opponents attacked him in the primary debates] emerged in candidate advertising. Initially, almost all the attack ads – those ads criticizing a candidate or criticizing a candidate while promoting the candidate sponsoring the ad – focused on candidates other than Trump (figure 4.5). It was, again, only after Trump’s victories in South Carolina and Nevada that Trump’s opponents – mainly Rubio and Cruz and their allied super PACs – began attacking Trump via paid advertising. This was yet another way in which Trump had a long, easy ride.”
I regard this as the most important chapter in the book. It poses, and answers, the following key question.
“But which voters ultimately came to support him, and why?”
After pointing out that Trump was an unlikely Republican candidate, who had previously held many non-Republican positions, the authors look at Trump’s three central campaign themes:
The authors remind us that political campaigns can affect the criteria voters used to make decisions. This is known as “political activation.” Because a candidate campaigns on certain issues, that campaigning can make those issues more salient to voters.
Accordingly, the question becomes whether Trump’s racist campaigning made voters focus more on racial issues, or whether those racial attitudes already existed.
Here, there is valuable data, illustrating just why this is such an excellent book.
“Fortunately, a novel and unusual survey helps address this issue. In July 2016, a month after the primary’s conclusion, the Views of the Electorate Research (VOTER) Survey interviewed 8,000 respondents who were originally interviewed in 2011 – 12. This 2011 – 12 survey captured respondents’ views long before Trump’s presidential candidacy, meaning that these views could not have been affected by his rhetoric in the 2016 campaign. In July 2016, this survey then asked Republican primary voters which of four Republican primary candidates – Trump, Ted Cruz, John Kasich, and Marco Rubio – they had supported. This survey thus identified whether Republican voters’ opinions measured well before 2016 were associated with support for Trump in the primary – and which opinions appeared to matter most.”
Below is figure 5.5 from the book. These are attitudes measured in 2011, long before Trump’s campaign, and therefore not influenced by him.
The first quadrant looks at the views about the reasons for racial inequality. Voters who believe that African-Americans are poor, not because of discrimination, but because they lack effort were far more likely to support Trump.
Similarly, Trump supporters had less favourable views of African-Americans, less favourable views of immigration and less favourable views of Muslims.
The rest of the chapter demonstrates with additional details the extent to which racial anxiety motivated voters to support Trump.
This chapter looks at the Democrat party primary contest between Bernie Sanders and Hillary Clinton.
The authors point out that the data shows declining support for Clinton amongst independent leaning Democrat voters and that she would have a significant challenge ensuring that Sanders’ voters would turn out for her in November in adequate numbers.
This chapter looks at the campaign between Trump and Clinton during the summer, again with detailed data on the extent of media coverage for the candidates, and voters fluctuating views of them.
The key late development in the campaign was of course the letter from FBI Director Comey to Congress on 28 October announcing that he had reopened the investigation into Hillary Clinton’s emails. The authors look hard at the data to try to assess what impact this had on the final result. They write “Moreover, there is no clear evidence that the Comey letter affected people’s intention to vote for Clinton.”
However, they then go on to write:
“Of course, it is impossible to know for sure whether the election would have turned out differently without Comey’s letter. It is simply important to acknowledge the uncertainty underlying any such hypothetical.
But focusing on the Comey letter may have the unintended consequence of underestimating the effect of Clinton’s controversies. The coverage of Clinton’s scandals was not only more extensive than coverage of Trump’s scandals but arguably created a more coherent narrative, with ongoing news coverage and revelations that emerged regularly over time. By contrast, Trump’s scandals tended to come and go quickly and concerned many different topics, including his business practices, controversial positions on issues, and behaviour toward women and minorities. Thus, the overall impact of Clinton’s scandals and news coverage of them was bigger than the impact of the Comey letter itself.”
The authors go on to discuss how Clinton’s long and often hostile relationship with the news media made it harder for her to dispel the image that Trump was promoting of her being dishonest secretive and unethical.
Trump won the Electoral College, and therefore the presidency, by 304 – 227 despite Clinton winning the popular vote by 2.1%. The authors write: “The extraordinary divergence between the popular vote and the Electoral College vote means that there is no simple way to explain or interpret the election outcome.”
They devote the chapter to a detailed analysis of the result and which categories of voters voted in particular ways.
This chapter takes the story into the period of Trump’s presidency. It looks at how identity politics have intensified under President Trump. The most fundamental question is this: "Is American identity about race or about a set of ideas?"
“In the December 2016 VOTER Survey, more Americans indicated that American identity is about beliefs, not race or religion.…
But the seeds of division about American identity were also apparent in this survey. Republicans and Democrats did not always agree – particularly regarding the importance of being Christian to being American. A majority of Republicans (56%) said that being Christian was very or fairly important, compared to 30% of Democrats.”
The authors concluding paragraphs are about the choices that America faces in the future, which we are now seeing in 2020:
“What gave us the 2016 election, then, was not changes among voters. It was changes in the candidates. Only four years earlier, issues like race and immigration were not as central either to the candidates or to voters. The 2016 election was different because of what the candidates chose to do and say – and then, after the election, because of what Trump has chosen to do and say as president. Those choices have consequences for voters.
Political leaders will always have those choices. They can call someone un-American or a “son of a bitch” or “deplorable.” They can call someone’s country a “shithole.” They can tell us to “beat the crap” out of someone they disagree with. They can also ask us to welcome others, to find common ground, and even to heal the country. These choices are what helped build the identity crisis in American politics. They are also what can help take it apart.”
Because America is the most powerful country in the world, the outcome of the American presidential election is always important.
The election of Mr Trump came as a surprise to almost everyone. This book provides by far the most coherent and best evidenced example I have seen of why Mr Trump won.
Reading it is vital preparation for looking forward to the November 2020 election.