FiveThirtyEight

Saturday was Joe Biden’s first-ever win in a presidential primary or caucus. It was an awfully big one: Biden won South Carolina by nearly 30 percentage points over Bernie Sanders. And it made for one heck of a comeback: Biden’s lead over Sanders had fallen to as little as 2 to 3 percentage points in our South Carolina polling average in the immediate aftermath of New Hampshire.

What explains the big swing back to Biden in South Carolina? And what does it mean for the rest of the race — and in particular for Sanders, who had entered this weekend as the frontrunner?

Here are five possible explanations — ranging from the most benign for Sanders to the most troubling for his campaign.

Hypothesis No. 1: This was a “dead cat bounce” for Biden because voters were sympathetic to him in one of his best states. It may have been a one-off occurrence.

Remember Hillary Clinton in New Hampshire in 2008? Left for dead by the national media after she lost Iowa to Barack Obama in 2008, she overcame a big polling deficit for an upset win in the Granite State. It didn’t do her much good, though; she won Nevada the next week but badly lost South Carolina two weeks later, eventually losing the nomination to Obama.

There are some similarities to Biden’s position in South Carolina. Like Clinton before New Hampshire, the media all but counted him out of the running after Iowa. Like Clinton in New Hampshire, Biden had a strong debate a few days before the primary along with some emotional moments on the campaign trail. Furthermore, some of the reporting from South Carolina suggests that certain South Carolina voters — especially older whites and African-Americans — felt deep loyalty toward Biden and wanted to keep him in the running.

Degree of concern for Sanders if this hypothesis is true: Low to moderate. If this were truly just a one-off sympathy bounce, then Sanders can live with it. Sure, Bernie missed an opportunity to put the race away with a win — or perhaps even a close second — in South Carolina. But voters rarely just hand the nomination to you without creating a little bit of friction. But if voters in other Super Tuesday states feel the same way that South Carolinians did, the sympathetic moment for Biden may not be over yet.

Hypothesis No. 2: The disparate results so far are simply reflective of the geographic and demographic strengths and weaknesses of the candidates. The notion of “momentum” is mostly a mirage.

If this is the case, you could wind up with a very regionally-driven primary, with Biden doing well in the South but perhaps not so well everywhere else. This is more or less what our model expects to happen, for what it’s worth; it now has Biden favored in every Southern Super Tuesday state except Texas, and he’s an underdog everywhere outside of the South.

The counter to this: Biden clearly did much better in South Carolina counties and precincts that weren’t as emblematic of his base than he had in those kinds of districts in other states. The counter to the counter: Geographic factors pick up a lot of information that demographics alone miss. So his strong performance in certain parts of South Carolina may bode well for how he’ll do in Alabama or North Carolina or Georgia. It may not say much about his performance in Michigan or California, however.

Degree of concern for Sanders if this hypothesis is true: Low to moderate. Sanders led Biden by about 12 points in national polls heading into South Carolina. Moreover, our model — which uses demographics in its forecast — has Sanders ahead. So although Biden has some strong groups and regions, Sanders’s coalition looks as though it’s slightly bigger and broader overall — although a post-South Carolina bounce for Biden or swoon for Sanders could eat into that advantage.

Hypothesis No. 3: The party is finally getting behind Biden. It may or may not work.

Almost half of South Carolina primary voters said that Rep. James Clyburn’s endorsement of Biden was a big factor in their decision. There are some questions about the cause and effect: It may be that Biden voters were pleased with the endorsement and said it was a major factor, even though they were planning to vote for Biden already. Still, Biden did get a big, late surge in the polls following the debate and the endorsement.

Clyburn is also one of the few party bigwigs to have endorsed a candidate. While lots of U.S. representatives, mayors, lieutenant governors and so on have endorsed, not many senators, governors or party leaders have. That leaves open the possibility there could be a surge of endorsements for Biden in the coming days. He’s already scored several major endorsements in Virginia, for instance, which is a Super Tuesday state.

Degree of concern for Sanders if this hypothesis is true: Moderate. The “Party Decides” view of the race treats endorsements and other cues from party leaders as being highly predictive and important. And a surge of endorsements for Biden seems reasonably likely. This could reverse a longstanding period of seeming indifference by party leaders toward Biden as they hoped for Michael Bloomberg or some other alternative to emerge.

But it’s not clear how effective an endorsement surge would be, as few legislators command the respect in their states that Clyburn does. Moreover, although we’re not going to cover it at length here, there’s plenty of room to question how empirically accurate the “Party Decides” is. Meanwhile, endorsements aren’t necessarily what Biden needs; an influx of cash would do him more good.

Hypothesis No. 4: Voters are behaving tactically. Biden was the only real alternative to Sanders in South Carolina, and he may be the only real alternative going forward.

Tactical voting is something you hear a lot about in multi-party systems like the United Kingdom’s, where voters are trying to find the most viable candidate from a number of similar alternatives (for example, from among the various parties that opposed Brexit). The same dynamics potentially hold in multi-candidate presidential primaries, and we’ve already seen evidence of it. In New Hampshire, voters flocked to Pete Buttigieg and Amy Klobuchar in the closing days of the campaign and away from Biden and Elizabeth Warren. In South Carolina, tactical voting may have worked in Biden’s favor, instead. Biden was fairly clearly the most viable alternative to Sanders, so voters for candidates like Tom Steyer and Buttigieg may have gravitated toward him in the closing days of the campaign.

Degree of concern for Sanders if this hypothesis is true: High. First, if voters are actively looking for alternatives to Sanders — but just can’t settle on which one is best — that can’t be good news for him, and gives some credence to the “lanes” theory of the race in which the moderate vote could eventually consolidate behind one alternative to Sanders. The South Carolina exit poll had Sanders’s favorability rating at just 51 percent, which is some of the stronger evidence for a ceiling on his support so far.

Moreover, Biden’s strong finish in South Carolina, along with improved debate performances, endorsements, and increasingly favorable media coverage, could make it clear to voters that Biden is the best alternative to Sanders after all, possibly with some exceptions where there are home-state alternatives (Klobuchar in Minnesota and Warren in Massachusetts). If Biden picks up support from tactical voters who had previously backed candidates such as Bloomberg and Buttigieg in polls, that could lead to a larger-than-usual South Carolina bounce.

Hypothesis No. 5: There has already been a national surge toward Biden that is not fully reflected in the polls.

It didn’t get much notice, but polling outside of South Carolina was also pretty favorable to Biden toward the end of last week, including polls that showed sharp improvements for him in states such as Florida and North Carolina. He’s also gotten better results in some national polls lately — climbing back into the low 20s — along with other, not-so-great ones.

The data isn’t comprehensive enough to know for sure. Between the dense cluster of events on the campaign trail (primaries, debates, etc.) and the different races that pollsters are surveying (South Carolina, Super Tuesday, national polls), everything is getting sliced pretty thin. But we do know that Biden made big improvements since the debate in South Carolina polling, the one state where we did have enough data to detect robust trendlines.

Degree of concern for Sanders if this hypothesis is true: High. Suppose that Biden gained 5 or 6 percentage points across the board nationally and in Super Tuesday states as a result of this week’s debate (or other recent factors such as voters’ reaction to coronavirus), but it’s gone largely undetected because there hasn’t been enough polling. If that’s the case, then Biden may already be in a considerably better position than current polling averages and models imply — and then he could get a further bounce from winning South Carolina on top of it. This is a scary possibility for Sanders, and although there isn’t enough data to prove it, there also isn’t much that would rule it out.


Filed under