About 12 percent of Bernie Sanders supporters from the Democratic primary crossed party lines and voted for Donald Trump in the general election, a new analysis says.
In several key states — Pennsylvania, Wisconsin, and Michigan — the number of Sanders to Trump defectors were greater than Trump’s margin of victory, according to new numbers released Wednesday by UMass professor Brian Schaffner.
Sanders -> Trump voters…— G Elliott Morris (@gelliottmorris) August 23, 2017
Trump win margin…
In an interview, Schaffner noted that in an election this close, any number of voting blocs could have proved decisive. And the analysis certainly doesn’t necessarily prove Sanders would have won — Schaffner also found that 34 percent of John Kasich’s GOP primary supporters backed Clinton in the general; perhaps more would have stayed in the Republican camp had Sanders been the Democrats’ nominee, or perhaps fewer of Hillary Clinton’s voters would have voted for Sanders. Then again, it also suggests some voters were in Sanders’s reach that were out of Clinton’s.
Moreover, defections from a primary to general election are common. More voters went from Hillary Clinton to John McCain in 2008 than went from Sanders to Trump in 2016; about 13 percent of Trump’s 2016 voters also voted for Barack Obama in 2012.
“The way to think about this is, as several people have noted, that this election was so close that any number of things could have proved the decisive difference,” Schaffner said in an interview. “This is yet another one of those anythings.”
But given Democrats’ interest in winning back the Rust Belt, it’s worth digging into exactly who this population of voters is. Schaffner found some demographic characteristics that might align with what you’d expect — Bernie-Trump voters were older and whiter than the average Democratic primary voter, for instance.
Perhaps surprisingly, however, these defectors did not turn out to have views on trade policy that marked them as significantly more opposed to free trade than the average Democrat. That may fly against the expectation that Sanders’ views on trade were unique to his appeal, but some political scientists were making that case as early as April 2016.
Also of note: the Bernie-Trump voter also proved much more likely to consider himself or herself “somewhat conservative” or “very conservative” than the average Democrat. Sanders, of course, ran on a policy platform well to Clinton’s left — but was able to do so in a way that allowed him to win over voters that disdain the “liberal” label.
A transcript of my conversation with Schaffner on Wednesday follows.
I’ve seen some people who are using your new finding to trumpet that “Bernie Would Have Won,” because the number of Bernie-Trump voters is greater than the margin of difference between Clinton and Trump. And then there are others who say that this poll proves that Bernie voters cost Hillary the election, because his supporters stubbornly flipped to Trump in the end.
And it seems like for one claim to be correct, so does the other one.
One piece of this that’s important to keep in context is that you always see this kind of defection between a primary and a general election. In 2008, you saw a lot of Hillary Clinton voters who ended up backing John McCain — so it's not abnormal to see this kind of thing. And more of them did so in 2008 than this time. [15 percent of Clinton’s 2008 voters in the primary supported McCain in that year’s general election.] Although given the candidates this time versus in 2008, it may have been surprising to see even this rate of defection.
The thing that really stood out to me is that a lot of these people who voted for Sanders — and then Trump — don't look like modern day Democrats. So you saw a lot fewer of them actually identify as Democrats than your normal Sanders voter; and, even more striking, they seem to have views on racial issues that are far more conservative than your typical Democrat.
But there does appear to be a racial component to this, as defectors are much more likely to disagree that whites are advantaged in US 9/n pic.twitter.com/zcoHm9APNf— Brian Schaffner (@b_schaffner) August 23, 2017
It's not clear to me this necessarily if this is a Hillary Clinton problem specifically, or if this was Bernie Sanders having a special appeal to bring people into Democratic Party primaries, who would otherwise be inclined to be Republicans.
So who are these Bernie-Trump defectors?
I assume they’re not the young voters who made up the core of Bernie’s base in the primary? This is the other major part of his coalition — the white union “Reagan Democrat” voters.
But then again, you have a chart showing that these Bernie-Trump defectors weren’t particularly opposed to free trade deals:
For those suggesting this is about trade, note that opposition to TPP is not dramatically different among defectors. 8/n pic.twitter.com/Fu0GhL58bU— Brian Schaffner (@b_schaffner) August 23, 2017
That’s a good question. I’ll just say I haven’t gone down every rabbit hole yet. But they were also older and whiter [than the average Bernie voter in the primary], and also less liberal.
Of the ones that switched to Trump, only about 25 percent also voted for a Democratic candidate for Congress. And we do have a little bit about what they did in 2012 — it looks like they were split roughly evenly, 50-50, between Obama and Romney. So these appear to be people who are trending out of the Democratic Party.
Of those Bernie voters who supported Trump in the general election, the average age was 52. Those who stuck with Clinton were an average age of 45, and of those who broke for a third party, the average age was 44. Of those that didn’t vote, their average age was 35 — these were the ones that got activated by Bernie, and then dropped back out when he didn't win. It’s worth noting that very few of the primary voters stayed home. People who vote in primaries are highly engaged in politics — they’re not people who come in and out of the electorate.
Some asked for more detail on how Sanders primary voters behaved in general. This graphic shows this, including small % who abstained 2/n pic.twitter.com/iOjKr7eoYJ— Brian Schaffner (@b_schaffner) August 23, 2017
I also looked at how the Bernie-Trump voters identify themselves on the ideological scale, and very few say that they're liberal. Only about 17 to 18 percent say that they're liberal, in any kind of way, shape, or form, though they voted for Sanders.
By contrast, about 45 percent of these Bernie-Trump voters say they're ‘middle of the road’ — basically, a lot of them see themselves as “moderates.” Meanwhile, another 35 percent of them are claiming to be either somewhat conservative or very conservative.
I think what this starts to suggest to me is that these are old holdovers from the Democratic Party that are conservative on race issues. And while Bernie wasn't campaigning on that kind of thing, Clinton was much more forthright about courting the votes of minorities — and maybe that offended them, and then eventually pushed them out and toward Trump.
How big or decisive was this population?
The way to think about this is, as several people have noted, that this election was so close that any number of things could have proved the decisive difference. This is yet another one of those anythings.
An 8 to 10 percent defection rate among Sanders voters changes the outcome in Pennsylvania, Wisconsin, Michigan — even maybe some other states.
But at the same time you have to remember how many Republican primary voters didn’t end up voting for Trump. It’s probably not quite as many, but it’s more of a wash in that respect.
It’s been months since the election — why is this data coming out now?
This is the Cooperative Congressional Elections Study data, and we’ve had the survey data itself since February. And I saw this pattern back in February, but I was a little skeptical.
What we do with this survey is go back and match respondents to voter files because people lie about having voted, and especially lie about having voted in the primaries. What I was worried about is people who claimed to have voted in a primary for Bernie but didn’t.
So I waited until I had the data that matched the respondents to their voter file record, which allowed us to see — these are the people we know voted for a primary. And we’re still getting this defection rate.