Sabato's Crystal Ball

Did Bernie Sanders Cost Hillary Clinton the Presidency?

Robert Wheel, Guest Columnist October 12th, 2017

facebooktwittergoogle_plusredditpinterestmail

After months out of the limelight, Hillary Clinton edged back into view recently with two fits of activity. The first was an announcement that her voters should read Verrit, a website managed by a former Clinton digital strategist that purports to post verified facts for the 65.8 million people who voted for her. One of the site’s first such facts was that Bernie Sanders helped put Donald Trump in the White House. Later on, an excerpt from Clinton’s new book leaked, in which she blames Sanders for hobbling her in the general election, though she seems far more circumspect about why she lost in general. Still, this all begs the question, did Bernie Sanders really put Donald Trump in the White House?

To answer that question, first we need to acknowledge the limitations of such an inquiry. Individual presidential elections have an n of 1; there’s no control group in which there’s an election in which Clinton glides through a primary unscathed. Accordingly, one cannot definitively say “but for one event, another outcome would have occurred.” Especially one that’s so hard to quantify. The Comey Effect can at least be measured to some degree because it occurred when the race was being polled daily, but even then it’s impossible to isolate from a universe in which there was no letter sent to Congress by then-FBI Director James Comey just days before the election.

Moreover, Trump’s margin was slim, but not so slim that we can attribute myriad exogenous factors to his victory. Al Gore’s narrow loss of 537 raw votes in Florida in 2000 can be attributed to any number of small factors (the Supreme Court, Florida voter purges, the butterfly ballot, Katherine Harris, Ralph Nader, his choice of running mate) in addition to the larger issues that plagued his campaign. Clinton’s aggregate raw vote loss in Pennsylvania, Wisconsin and Michigan was around 78,000 votes, closer to John Kerry’s 119,000 raw vote loss in 2004 (if he had flipped Ohio he would have won) than to Gore’s. Accordingly, while there are still numerous events that could have changed the outcome of the election, it’s harder to say definitively whether they actually did.

Before we turn to the Sanders effect, we should establish that Clinton was the favorite heading into the general election. In 2008 and 2012 Barack Obama won with potentially durable majorities, and all Clinton really needed to do was get the same voters to back a Democratic candidate for a third time. However, faced with the prospect of a crass and corrupt Republican nominee, she tried to broaden the Democratic electorate as much as possible instead of trying to consolidate Obama’s base, ignoring states in the “Blue Wall” like Michigan and Wisconsin and diverting resources to areas she didn’t really need to win like Arizona and Georgia. Accordingly, I examine “did Sanders cost her the election” through the prism of whether he cost her votes from the Obama coalition that she was unable to fully replicate.

Next, let’s turn to the Sanders effect. It makes sense to look at how his behavior affected (if at all) the following groups:

Fortunately, the Cooperative Congressional Election Study has collected data that can help guide us in determining whether they would have voted for Clinton if not for the Sanders campaign.

Sanders-Trump voters

The phenomenon of Sanders-Trump voters is perhaps the easiest to dismiss out of hand as costing Clinton the election. Much of the phenomenon is attributable to the function of party registration. Registration with a political party is a lagging indicator. For example, West Virginia was Trump’s second-best state in terms of vote percentage and percentage margin (after Wyoming). However, registered Democrats still outnumber registered Republicans in the Mountain State. That’s because so many West Virginians registered as Democrats when the state was still solidly blue and never changed their registration to Republican as they started voting differently. Switching party registration involves paperwork and deadlines, so less involved voters don’t necessarily get around to making a change. So in closed or semi-closed primary states like Oklahoma, Kentucky, and West Virginia, these ancestral, less-engaged Democrats can’t vote in Republican primaries. In 2008, they voted overwhelmingly for Clinton, who was running to the right of eventual-winner Barack Obama. In 2012, these states gave outsized numbers of votes to barely-funded, fringe Democrats running against Obama. And in 2016, Oklahoma and West Virginia thumbed their nose at Clinton and voted for Sanders while Kentucky backed Clinton but by less than a point.

The data bear this out. The Sanders-Trump voters didn’t self-identify much as Democrats, had approval ratings for President Obama of 23% (perhaps he’d be a better scapegoat for Clinton among this subset of voters) and were less likely to believe that white people have certain advantages in the United States. This doesn’t sound like people who were part of the Obama coalition, nor would they have backed Clinton against Trump in most instances. To be clear, there are some people in this group of voters who might have voted for Clinton if not for the Sanders campaign, but they are the vast minority, and you should be wary of anyone arguing otherwise.

Sanders-third party voters

Of course, Bernie Sanders’ voters didn’t need to vote for Trump to help out the eventual president. Those who voted for third parties did support Obama a decent amount: 67% of these voters who said they cast a 2012 vote said they voted for Obama in 2012 (though that number probably overstates his support among this cohort — election winners traditionally poll better after their victories) and 59% of all such voters approved of Obama’s job performance. But only 10% of these voters called themselves Strong Democrats (25% overall identified as Democrats). Accordingly, while this group was part of the Obama coalition, it was a pretty weak one. Demographically the group was pretty ordinary; close to the national average in gender identity, race, and family income. They were actually slightly less likely to self-identify as liberal than other Sanders voters, suggesting the group is perhaps a bit more contrarian than ideological. In any event, these voters are not the stereotypical “Bernie Bros” — strident, young male liberals who frequently post online.[1] Rather, this is a group that leaned Democratic but was never reliably part of the base. In other words, this group was always going to be the least likely of Sanders’ base to stay with Clinton. There’s little evidence that he turned these people away from voting for President Obama to voting for President Clinton.

Sanders-nonvoters

Unfortunately we just don’t have enough data on the people who voted for Bernie Sanders in the primary but then sat out the general election. When working with such a small sample size it’s difficult to extrapolate conclusions about the population. However, I will try to address the behavior of his voters, including those who did not vote, in the following section.

Sanders effect on his own voters

From the data above, I don’t think you can make a convincing case that Sanders cost Clinton the election based on how his own voters behaved. A higher percentage of his voters backed Clinton than her voters backed Obama in 2008 or Rubio and Kasich voters backed Trump in 2016. Assuming that his voters cost her the election ignores the fact that, if he had not run, in all likelihood there would have been another credible Democrat that ran against Clinton.

Every non-incumbent presidential nominee has faced a contested primary since Richard Nixon cruised to the Republican nomination in 1960, back in an era where convention delegates were still largely chosen by state parties and not directly by voters. Once voters started electing convention delegates, even sitting vice presidents George H.W. Bush and Al Gore faced competitive primaries. In April 2015, a Gallup poll indicated that Democrats wanted Clinton to be their party’s nominee by a 57%-38% margin; that is a strong level of support, but also not one  that heralded what might have amounted to nomination-by-proclamation.

More than a third of Democrats, 38%, did not want Clinton to be the nominee, so if Sanders hadn’t filled the oppositional role, there would have been some other candidate that did. Maybe Martin O’Malley could have successfully positioned himself as the Clinton alternative. Maybe Elizabeth Warren or Joe Biden would have reconsidered after seeing a weak non-Clinton primary field develop. Maybe one of Sanders’ congressional backers like Reps. Tulsi Gabbard of Hawaii or Keith Ellison of Minnesota would have decided to take the plunge, or a liberal celebrity like George Clooney or Mark Cuban would’ve seen Trump leading Republican primary polls and figure “why not me?”

In any event, it’s difficult to tell if another candidate would have done better or worse than Sanders, especially because Sanders could have done better. He got 43% of primary/caucus votes so he outperformed a generic non-Clinton candidate, but not by much. In the 2008 cycle, Obama entered the race in early 2007 when Clinton was also polling well (although her level of support was smaller nationally than it was in 2015), and he managed to win the Democratic nomination.

Moreover, when Sanders said “The American people are sick and tired of hearing about your damn emails” at the first primary debate, he unilaterally disarmed himself of one of the strongest arguments against Clinton: that her (and her husband’s) history of dwelling in that gray zone between legality and impropriety made her vulnerable in the general election. Sanders might have won if he hammered home the message that Clinton was unelectable. Based on the CCES data, 36% of her primary voters described themselves as liberal and 9% as very liberal — surely some of them were voting for a candidate that they thought was most likely to win in November than the one they best aligned with politically.

Additionally, Sanders’ post-election activities indicate that he is not a very savvy organizer. The candidates that he’s endorsed have not performed well and his voter mobilization organization is frequently involved in People’s Front of Judea/Judean People’s Front squabbling. A more organized candidate might have performed better against Clinton.

Further, Clinton’s favorability among Democrats in October/November 2016 was essentially the same as her favorability in April 2015. The numbers were slightly lower, but it’d be difficult to attribute any such drop directly to Sanders considering everything else transpiring between April 2015 and the 2016 general election. And if there was a drop arising out of the primary campaign, it could have been worse if another candidate had been her main adversary.

Accordingly, the idea that Sanders cost Clinton the presidency because of his own voters’ behavior simply isn’t very compelling. His voters turned out relatively well for her and there’s little proof that those who didn’t were members of the Obama coalition who would not have voted for Clinton but for Sanders.

Sanders effect on all voters

Of course, there’s also the argument that Sanders undermined Clinton among all voters, not just her own. The argument has numerous prongs:

The first allegation is easy to disprove because it assumes, as do other arguments, that the non-Sanders universe did not have a vigorous primary challenge of Clinton. Sanders admitted that after the April 26 primaries that he was mathematically eliminated (the Democratic Party practice of allowing formally unpledged superdelegates to vote for a nominee makes such determinations inexact) and was only staying in the race to influence the party platform, eventually dropping out and endorsing Clinton two months later, a couple of weeks after she clinched a majority of delegates. This is typical behavior for eliminated candidates:

So, Sanders behaved exactly as many other eliminated primary challengers, including Clinton herself. Blaming a loss on normal behavior is disingenuous and, in this case, hypocritical.

The second allegation is also easy to disprove. Sanders fully endorsed Clinton at the convention. He campaigned for her regularly and told his supporters not to support third parties. Contrast this behavior to Cruz, who in a primetime convention speech told his supporters to vote their conscience (a rebuke of Trump that led to boos in the convention hall) while Kasich didn’t attend the convention, never endorsed Trump, and wrote in John McCain for his vote. Trump has a far better case that Kasich and Cruz let Clinton get too close than Clinton has a case that Sanders cost her the presidency.

The third allegation is more serious than the other two, so it requires a bit of unpacking. In July 2016, WikiLeaks published internal DNC emails disparaging Bernie Sanders and his supporters, asking if there was a way to thwart him in the Kentucky and West Virginia primaries, calling him a liar, and generally being dismissive of his campaign. However, Sanders never walked back his support of Clinton. The blame here really belongs on a) the parties behind the disclosure and b) the inept leadership of the DNC, led by Debbie Wasserman Schultz.

Moreover, Sanders was never going to be the reason people attacked Clinton as crooked. She has been a national figure since the early ‘90s, and her husband’s administration was constantly hounded by investigations and people in and adjacent to it going to jail. Republicans had known for decades that the best way to go after Clinton was by attacking her ethics. And Clinton indulged these critics by engaging in dubious if not illegal behavior like using a private email server and soliciting donations from foreign governments for the Clinton Foundation. Ultimately these decisions were her own, knowing that if she did run for president again she’d be attacked like her husband was, and had nothing to do with Sanders.

So I don’t see any compelling reason to think that Sanders somehow cost Clinton the election among people who didn’t vote for him. He behaved no different than past defeated candidates for nomination (and was certainly more supportive than the candidates that Trump beat) and refused to turn himself into a victim after the DNC hack.

Conclusion

Turning to the inquiry of this piece, did Bernie Sanders cause Hillary Clinton to lose the presidency? I’m an attorney, so I think the best way to approach some inquiry is to determine whether he was the proximate cause of her loss. That is to say, if not for his candidacy, would she be president?

There are many proximate causes for Clinton’s loss, and I think you can divide them into three broad categories:

Certainly all of those are more compelling proximate causes than Sanders, but is Sanders himself a proximate cause? Again, I don’t think he is unless you assume that Clinton would have had no serious primary opposition. I don’t dispute the fact that he did end up costing her some votes. And her 78,000-vote loss was certainly slim enough that it can have multiple proximate causes. I’d be willing to state that Bernie Sanders might have cost Clinton the election if no serious primary challenger emerged in the counterfactual. In Pennsylvania, Clinton would have needed around a quarter of the Sanders voters who didn’t vote for her to have changed their votes to win. How many of those people were among the 20% of Democrats who didn’t approve of Clinton before Sanders even entered the race? My guess is most of those people weren’t going to vote for Clinton anyway, but I doubt we can ever be sure.

If you’re a Clinton voter angry at Sanders, is that more compelling than any of the reasons listed above? I don’t think so. And if you want Democrats to win back the presidency in 2020, wouldn’t it make more sense to dwell on issues that had a bigger impact, and focus on the ones that you can control going forward? I could understand relitigating the 2016 primary if it had caused serious damage, but the fighting appears to be based more on rehashing old grievances than actually winning voters back into the Democratic Party. Frankly, if there’s any Democrat other than Clinton deserving of Democratic ire, it’s President Obama. It was his neglect of the DNC (he’d been told that Debbie Wasserman Schultz should be ousted as early as 2012) that led to widespread grassroots distrust of the party. It was his naiveté that Republicans would actually consider voting for Merrick Garland that prevented him from nominating an African-American or Latino nominee to the court that could’ve made the Supreme Court a bigger issue for Democrats. And it was his administration that didn’t stop the Comey Letter or publicize Russian interference in the election. If you want to start an intraparty squabble, it’s those types of mistakes that the party should seek to avoid repeating.

But your real ire should be directed toward Clinton, who seems to have admitted in her book that ultimately she’s the reason that she lost, and her campaign, where some of her staff is still blaming Sanders for their bungling.  These people are blaming the media for turning people against their candidate and their primary opponent for not turning out enough voters. You know who else faced media that didn’t like him and a party base suspicious of him? Donald Trump! Moreover, what is the point of a campaign other than to improve media coverage of your own candidate and turn out your own voters? At some point they need to take responsibility for their own actions. And if they still think that it was Bernie Sanders’ responsibility to turn out enough voters for the Democratic nominee, then they should have done the honorable thing and let him win.

Robert Wheel (a pseudonym) is an attorney and University of Virginia graduate who lives in New York. He tweets @BobbyBigWheel.

Footnotes

1. Regarding Bernie Bros, I do not dispute that they exist (though it does appear some were actually fake social media accounts set up by Russians). However, they were simply not a big part of his coalition of voters. The reason they were so prominent is that men in their 20s spend the most time online and, I speak as a former man in his 20s, are the most strident people online. If you were a Clinton diehard, please take solace in the fact that they were vastly outnumbered by Sanders voters who were not mean to you personally. So, it bears reiterating, the election was not your Twitter mentions.