March 09, 2016
America’s Brexit: Lessons from the 2016 Presidential Election
By Keith Hodson
The 2016 United States presidential election was destined to be historic. The country would either elect its first female president, or its first president with no government or military experience. Many Americans also agree that this has been the most vitriolic presidential campaign season in their lifetimes. Donald Trump had pushed his way through a crowded GOP field during the primary election – often insulting and berating members of his own party as much as he decried the activities of Democrats. The primary season for Democrats was significantly less bruising but no less unique: a centrist Democrat who was former First Lady, Senator of New York, and Secretary of State against a socialist Democrat who only became a Democrat to be eligible to run in the primary.
Once Donald Trump and Hillary Clinton clinched their party’s nominations, the vitriol of Donald Trump’s primary campaign continued into the general election. He touted the successes of Vladimir Putin – a man who no one in the developed world thinks has been legitimately elected – while simultaneously claiming he would only be happy to accept the results of the United States election “If he won”. Videos were leaked of him bragging about being able to “Grab women by the pussy, whatever”. Sexual assault allegations mounted against him and contestants of Trump beauty pageants consistently reported inappropriate and degrading actions taken by Trump. Most ironically, he claimed that the election of Hillary Clinton would lead to a constitutional crisis while the United States Senate was neglecting to perform its basic duty of confirming a new Supreme Court justice. Additionally, this was the second election in 16 years in which the Electoral College has led to a President being elected after losing the national popular vote. The only other two times in American history in which this has occurred was in 1876 and 1888. Since the election, polls have shown a slight majority of Americans now approve of abolishing the Electoral College entirely – and amending the Constitution to do so. Thus, it appears the true constitutional crisis has been caused by Trump’s, and not Clinton’s, nomination.
Leading up to the election, the majority of Americans polled did not believe that Trump would be able to win the presidency. Notable blogs and election experts had the chance of Hillary winning between 85 and 90 percent. Almost all major news organizations – including NBC, ABC, New York Times, Washington Post – had Clinton winning easily by two to three points in national polls. It was thought Trump’s attacks on women, minority groups, members of the LGBTQ communities, and others would be too damning for Americans to vote for a President with such a nonexistent moral compass. In pre-election polls Trump was losing the African American vote by 80 points, the Hispanic vote by 36 points, women’s vote by 12 points. Though among white men the American public was staunchly in support of Trump – with a 21 point lead amongst whites and 12 point lead amongst men. His boasts of his lack of political experience, and his willingness to game the political and business systems with which he worked, meant many that Americans believed he wouldn’t have their best interests at heart, or that he wouldn’t be able to follow through on many of his campaign promises. Over 65 percent of Americans polled agreed that Trump would be “unfit for the Presidency” – the highest number in this history of such polls – while only 40 percent said the same about Clinton, though this was also a historically high number. In addition, a Washington Post poll demonstrated that 8 in 10 Americans thought paying taxes was a civic duty – this contrasts with Trump’s statement that not paying federal income taxes made him “smart”.
This political backdrop, and the stark differences with the political reality of the election outcome, has left many Americans asking “What went wrong?”. Trump and his supporters would likely claim that his campaign mobilized an entirely new voting block within the United States, and activated a new type of anti-establishment populism that was also seen in the Bernie Sanders’s campaign. However, this is simply not the case. Trump got the same number of votes in 2016 as Romney did in 2012 – approximately 60 million – with a similar overall voter turnout. What is the case, is that this election was a battle between two very unpopular candidates. Over 60 percent of Americans saw Trump unfavorably, with Clinton not far behind at between 55 and 60 percent unfavorable ratings. Even within their own parties the candidates were not very popular – with only 70 percent of Republicans supporting Trump and similar numbers of Democrats actively supporting Clinton. This year over six million votes were cast for third party or write-in candidates, as opposed to barely two million in 2012.
Additionally, the Republican Party and Democrat Party have shifted in very important ways in the last thirty years. For instance, in 1992 college education was not a strong indicator of party preference in presidential elections. Since then, college education has become an increasingly important indicator of party preference. This year Trump beat Clinton by 39 points among white non-college educated voters and beat Clinton by 4 points among white college educated voters. On the other hand, Clinton beat Trump by 21 points among those with some form of post-graduate education. The Democrat party has also lost the poor working class white vote that propelled the party during the last 50 years – especially evident in places like Michigan and Pennsylvania (states that last voted for a Republican almost 40 years ago), and even West Virginia (a state that voted Democrat for president every year from 1866 to 2000). These states have seen shifts of 20 percent or more among white non-college educated voters, i.e. low income working class voters, from the Democrat to Republican party. Clinton also failed to produce the types of coalitions among minority groups that Obama enjoyed in 2012 – with fewer Latinos and African-Americans voting for Clinton in 2016 than voted for Obama.
Beyond looking at shifts in the electorate (and the activation of the poor, uneducated, white America) it is important after this election to ask: how did the polls make so many mistakes? This is a question with much room for debate amongst political scientists, statisticians, and pundits but there is one theme that is clear: the technology of polling has not kept up with the technology of the times. Until very recently, landline cold calls were the primary mechanism by which the American public was contacted for polling reasons. Almost 9 in 10 households still had an active landline as of 2005 and the response rates for these polls meant that a reasonable assessment of public attitudes was possible. It was also important that a wide enough segment of the public used land lines to provide responses from a wide range of demographic groups. Naturally, as landline use has rapidly declined in the last decade so has the reliability of polls based on landline response. As of 2016 only 4 in 10 households have an active landline, and that number continues to decline. Additionally, only 17 percent of US households have just a landline for primary communication – meaning over half of those with a landline also have a cellphone.
While many polling groups have adjusted their polling methods to account for this loss of landline accessibility, and as such have added more cellphone responses to their reporting, flaws remain. For instance, response rates to poll calls was close to 80 percent in the 1970s, 40 percent in the 1990s, and now close to 10 percent in 2014. This means having to make more calls per question, which increases the price of calls and makes broad polling outside the scope of many organizations’ budgets. Not only this, but given the abundant use of caller id those who choose to answer calls for polls are systematically different than those who ignore such calls. This means that it is now nearly statistically impossible to get a representative sample of the American public using the same mechanisms that polling groups have relied upon for decades. Instead of direct responses, polling groups are then often left to perform complicated weighted calculations of their survey questions to attempt to impute what a representative sample of the American public would say if given the chance. The gross inaccuracies of these methods have been put drastically on display in this election, even more so than in 2014 – a year that saw significant error on the side of political pundits predicting election results.
While there is no clear response to the question of “What will a Trump presidency look like?”, there are some clear answers to how we may better understand the opinions of American voters. It will be important for public interest groups to use a wider range of tools to gauge public opinion: including social media and decentralized voter outreach programs. It is clear that a wide sector of the American public still feels separated from the daily activities of politicians in Washington, D.C. – an attitude that must be shifted if there is to be any rebuilt faith in the institutions of our national government. And it is clear that Americans live in a time more divided than any in recent memory on a wide range of issues, and that this division was shifted from the political realm to the personal. While it is unlikely that having an artifact of the new alt-right movement as President will help bridge these divides, it is necessary for the future of the United States that Americans find a way.