Updated 5:57 PM | Feb. 27, 2020

Election Update: Biden Surges In South Carolina

By Nathaniel Rakich

Filed under 2020 Election

After former Vice President Joe Biden finished second in the Nevada caucuses, someone at his post-election speech shouted out, “Comeback kid!” It seemed like an odd claim at the time — Biden finished more than 26 percentage points behind Sen. Bernie Sanders (going by county delegates). But based on the latest polls, Biden may indeed be experiencing a comeback in South Carolina, which votes on Saturday.

As Nevada went to the polls last Saturday, the FiveThirtyEight forecast considered Sanders the slight front-runner in South Carolina, with a 1 in 2 (46 percent) chance of winning it. It gave Biden a 2 in 5 (40 percent) chance. Since then, in a reminder of how fluid primaries can be, Biden’s chances have skyrocketed. As of Thursday at 5:38 p.m. Eastern (😉), Biden has a 14 in 15 (94 percent) chance of winning the Palmetto State, while Sanders’s odds are down to 1 in 20 (5 percent).

The first hint of Biden’s comeback came two days ago: Biden had only a small lead on Sanders in our South Carolina forecast at the time, but the first survey conducted entirely after Nevada, from Public Policy Polling, showed Biden up 15 points. At that point, the poll was an outlier. Well, not anymore. Since Wednesday, we’ve gotten six new polls of South Carolina, most of which have given Biden commanding leads.

• On behalf of the Charleston Post and Courier, Change Research polled the Palmetto State from Feb. 23 to 27 and found Biden with 28 percent, Sanders with 24 percent, Tom Steyer with 16 percent, Sen. Elizabeth Warren with 12 percent, former South Bend, Indiana, Mayor Pete Buttigieg with 11 percent, Rep. Tulsi Gabbard with 5 percent and Sen. Amy Klobuchar with 4 percent. However, Change Research has had big house effects this election cycle — their polls have tended to underestimate Biden and overestimate several other candidates, including Sanders, relative to other polls. Once adjusted for house effects by our model, this poll is actually more suggestive of a 12-point Biden lead (specifically, Biden 32 percent, Sanders 20 percent, Steyer 16 percent, Warren 9 percent, Buttigieg 7 percent).
• An Emerson College poll dated Feb. 26-27 gave Biden a strong lead over Sanders — and it was an even better poll for Biden considering Emerson’s Sanders-friendly house effect. The poll’s raw toplines were Biden 41 percent, Sanders 25 percent, Buttigieg and Steyer 11 percent each, Klobuchar 6 percent, Warren 5 percent and Gabbard 2 percent. But adjusted for house effects, the poll was more like Biden 40 percent, Sanders 20 percent.
• A Feb. 26 poll by Starboard Communications, a South Carolina political firm, gave Biden a huge lead: He had 40 percent support, Steyer had 12 percent, Sanders had 11 percent, Warren and Buttigieg had 9 percent each, Klobuchar had 6 percent and Gabbard had 2 percent. Stripping away the poll’s house effects only slightly reduces the size of that lead: Biden 38 percent, Sanders 13 percent.
• A Monmouth University poll conducted Feb. 23-25 put Biden at 36 percent, Sanders at 16 percent, Steyer at 15 percent, Warren at 8 percent, Buttigieg at 6 percent, Klobuchar at 4 percent and Gabbard at 1 percent. A top-notch pollster, Monmouth has no significant house effects, so this was a good poll for Biden, pure and simple.
• According to a Clemson University poll from Feb. 17-25, Biden had 35 percent support, Steyer 17 percent, Sanders 13 percent, Buttigieg and Warren 8 percent each, Klobuchar 4 percent and Gabbard 2 percent. This poll also did not need to be significantly adjusted for house effects.
• Finally, a Feb. 23-24 survey by East Carolina University yielded the following results, which were mostly unaffected by house effects: Biden 31 percent, Sanders 23 percent, Steyer 20 percent, Warren 8 percent, Buttigieg 6 percent, Gabbard and Klobuchar 2 percent each. Interestingly, though, this poll wasn’t any better for Biden than ECU’s South Carolina poll from two weeks ago, which gave him an 8-point lead.

It’s tempting to chalk up Biden’s comeback to his performance in Tuesday’s debate or his Wednesday endorsement by Rep. Jim Clyburn, who carries a lot of weight in South Carolina Democratic politics. But in reality, the polls above were mostly conducted before either of those two events. Instead, the dividing line seems to be the Nevada caucuses. In six South Carolina polls conducted between New Hampshire 1 and Nevada, Biden averaged 26 percent and Sanders averaged 22 percent. In the six polls conducted entirely since Nevada, 2 Biden has averaged 35 percent and Sanders has averaged 20 percent.

It’s certainly odd that Biden, and not Sanders, would have gotten a bump out of a state where Sanders won nearly twice as many raw votes, but that’s what it looks like. Perhaps it is the manifestation of establishment backlash against the suddenly-real prospect of Sanders becoming the nominee.

And accordingly, that prospect is now getting less likely again. As FiveThirtyEight editor-in-chief Nate Silver wrote earlier today, Biden winning South Carolina — especially by a big margin — would put him in good position to win, or at least net many delegates from, the many Southern states that vote just three days later, on Super Tuesday. That could set the table for a drawn-out nomination fight between Sanders and Biden — one that could even lead to a contested convention.

Indeed, in our overall primary forecast, the chance that no one receives a majority of pledged delegates after all states and territories have voted is up near its all-time high of 1 in 2 (49 percent)! But although his outlook has worsened, Sanders retains decent odds to win a delegate majority: 1 in 3, or 32 percent. For instance, he’s still a better bet than Biden, although the former vice president’s chances are up to 1 in 6 (16 percent).

And increasingly, it is looking like those are the only three realistic outcomes of the 2020 primary calendar: All other candidates, including former New York City Mayor Michael Bloomberg, have a 2 percent chance or less.