A poll released by CNN asking Democratic voters to express their preferences on potential Democratic presidential candidates shows former Vice President Joe Biden leading a potentially very crowded field. Notably, in this poll, Biden performs considerably better than Senator Bernie Sanders (33% to 13%), while all other possible contenders – including Senators Kamala Harris, Elizabeth Warren, Cory Booker, Kirsten Gillibrand, and Amy Klobuchar, and outsider candidates like former New York mayor Michael Bloomberg or celebrity lawyer Michael Avenatti – are in single digits.
The first, and most obvious, response to this poll is that it is very early to say much of anything about the 2020 presidential primary race. With the 2018 midterms a few weeks away, we are likely to see some activity among the different likely candidates by the end of the year, but if recent years are an indication we probably will not see most campaign announcements until the spring or even summer of 2019 Additionally, primaries are complicated elections to predict. In general elections, many voters rely on their partisanship to determine who to support. That is, someone who self-identifies as a Republican will nearly always vote for the Republican candidate. In primaries, this doesn’t help any of the candidates since all of them are from the same party. Thus, voters have to do a lot more work to figure out who to vote for. Combined, prognosticating on presidential primaries this far ahead is something of a fool’s errand.
That being said, let us dramatically overanalyze what the CNN poll, and political science research on presidential primaries, suggests about the way the 2020 primaries could play out!
How presidential candidates win presidential nominations in the modern primaries era has been the basis of a number of political science studies over the years. In this regard, the most discussed study by far has been the book The Party Decides by a group of scholars associated with UCLA. The Party Decides argues that leaders within the parties limit voter choice in the primaries by picking an anointed candidate during the invisible primary – the months during which presidential candidates are running campaigns but there are no primary elections yet. By providing this candidate with public endorsements, money, and qualified staff, party leaders are argued to increase the likelihood of the anointed candidate winning the primaries.
While it is clearly true that party leaders try to ensure that their parties’ voters pick presidential candidates they believe are good selections, the UCLA school theory of presidential primaries does have some limitations. On the most basic level, there is the problem of whether the support of party leaders automatically means that a candidate who wins their party’s nomination did so because of that support. That is, Hillary Clinton clearly was the preferred candidate of the Democratic party leadership in 2016, but did this support cause her to win the nomination? That is, Clinton – a former senator, secretary of state, and first lady with near universal name recognition – would still have been a formidable candidate even without any support from the party leadership. Second, as political scientist Wayne Steger has shown, many party leaders don’t actually participate in this process since they never endorse any candidates. If party leaders control the presidential nomination selection process (in part) through their endorsements, we might expect more of them to use that power.
Finally, while no political science theory is ever 100% correct, The Party Decides has a tough time explaining Donald Trump’s success in the 2016 primaries. Trump was very clearly not the choice of the Republican Party leadership, yet he won the primaries relatively easily. As I have argued elsewhere, part of the reason is that Trump was a ‘perfect storm’ candidate: party leaders might help make candidates more competitive by giving them a better standing – and therefore, more media coverage – in the race. However, Trump did not need party leaders to provide him with money or media attention, since he could provide plenty of both for himself. Trump certainly was a unique case, and very few people correctly predicted his success. Still, Trump’s nomination has to be considered a high profile failure of the UCLA theory.
There is another major alternative way of looking at the presidential primary process. This perspective – produced by political scientist Larry Bartels – relies on a more important role for voters in selecting the nominee. In his study of the 1984 primaries, in which former Vice President Walter Mondale won the Democratic presidential nomination after a strong challenge by Senator Gary Hart, Bartels argued that Mondale represented a somewhat obvious candidate to most Democratic voters. Indeed, Mondale had been on the national ticket in 1976 and 1980, and as the former vice president he had high name recognition.
That is not to say Mondale was everybody’s preferred candidate. Rather, Bartels argues voters fell into two camps: those who liked Mondale and those who did not like him. The voters in the former group were unsurprisingly pretty consistent in their preferences: they supported Mondale throughout the primary process. Those in the latter group were far less consistent in their preferences: rather than back one candidate, they shopped around among Mondale’s challengers. Crucially, Bartels argues these voters based their support for their preferred ‘non-Mondale’ candidate on how likely they believed those candidates to be to actually win the nomination, and that voters updated this understanding as the race continued.
The Bartels-perspective explains a number of presidential primary races in both parties in the past few decades. Additionally, it also shows the benefits and potential limitations of being the ‘obvious’ candidate. For example, Mondale faced a tough fight in 1984, but managed to win the nomination. However, while Hillary Clinton was the ‘obvious’ candidate in 2008, the Democratic voters who did not want her to be nominee outnumbered those who did. Initially, these not-Clinton voters were divided across the many other candidates running for the presidency in 2007-2008. But once Barack Obama’s candidacy began to catch fire, and his probability of winning the nomination increased, the not-Clinton voters united behind him, and he won the nomination.
Can this perspective on presidential primary races help us make sense of the CNN poll? Keeping in mind that any single poll provides limited information, the Bartels perspective of the way primaries work might suggest a way to think about what Biden’s current lead tells us about how the 2020 race may play out.
Bernie Sanders produced a remarkably strong challenge against Hillary Clinton in 2016 – building up a major voting base, and tremendous name recognition. One might have expected that showing to catapult Sanders to the position of ‘obvious’ candidate dividing the Democratic electorate. Yet, while Sanders comes in second, he’s only slightly ahead of the rest of the pack (Senator Harris, with 9% of support). Instead, Biden looks to be the Mondale type candidate in the coming race.
Does this mean the 2020 primary is Biden’s race to lose? Not necessarily. At 33% support, Biden is a far from certain that he would win the primaries. Indeed, if this poll is an accurate measure of Democratic voters’ preferences, 67% of them prefer someone who is not Biden. If Biden runs, and if a substantial subset of those voters unite behind one of his challengers, Biden could end up losing the nomination much in the same way Hillary Clinton did in 2008.
But we still don’t even know if Biden will run. While he has a strong record, and as the most recent vice president is a plausible nominee, age certainly could be an issue (Biden is 75). What if Biden does not run? It is theoretically possible that a large part of Biden’s supporters would fall behind one of the other candidates. But if Biden’s voters do not have a consistent preferred runner up, there may not be an ‘obvious’ Mondale-esque candidate in the 2020 Democratic primaries. If so, the race might end up as a free for all between many candidates who all have relatively little early support and will have to battle each other to break through.