Bernie Sanders will win more pledged delegates than Hillary Clinton in the second half of the Democratic nominating season.
In fact, he'll almost certainly win more pledged delegates than Clinton in the final three and a half months of the primary season.
And virtually without question, he'll win more states than Clinton in these final three and a half months -- it's just a matter of how many more.
He'll also close out the primary season, it appears, beating Donald Trump by as much or (more often) substantially more than Clinton in nearly every national and battle-ground state poll taken.
Yet none of it is a surprise, even in the context of a race the media told us was essentially over a month ago.
In fact, everything that's happening now in the Clinton-Sanders race was predicted, long ago, by either Sanders himself or the hard data of this election season. Moreover, none of what's happening is a surprise to the politicos on the Clinton side, either; that's one reason they're working overtime to control and then shift the narrative from the inevitability of a major Sanders comeback. While it's still up in the air whether that comeback will be total or near-total, only by manipulating the narrative can the Clinton campaign keep Sanders at bay.
And that's why understanding that what's happening now is no more or less than what was readily predictable a year ago is crucial to understanding the current state of the Democratic primary race. This means unpacking not just the Clinton camp's transparent attempts to skew the media narrative, but also, and more importantly, the hard data behind a comeback that could end up being every bit as historic as Sanders supporters are now suggesting it will be.
So here are the four items every voter needs to have a handle on as we enter the vortex of nonsense the Clinton campaign sent spinning into the election season as soon as Sanders eliminated 22.5% of her delegate lead in just twelve hours of voting in three states.
1. Hillary Clinton's reversal of her position on super-delegates.
A year ago, when Bernie Sanders was at 4% in the polls, and Hillary was the presumptive Democratic nominee, and Party-selected super-delegates began lining up in droves to support Clinton's candidacy for President, the Clinton camp was very, very fond of super-delegates.
Super-delegates were announcing themselves by the hundreds before a single vote had been cast -- anywhere -- and without meeting with all of the Democratic candidates for President first. Indeed, super-delegates were announcing themselves by the hundreds before any of the Democratic candidates for President had made the case for their candidacy to either voters, the media, or even Party elders.
That was fine with the Clinton camp.
Then super-delegates began being reported by the national news media as though they were in some way obligated to the candidates they endorsed. The Democratic National Committee told the media to stop doing this, as super-delegates don't cast any votes until the nominating season is over, and -- as happened to Hillary Clinton herself in 2008 -- can change their minds up until seconds before they declare themselves at the Democratic National Convention. They can even --as happened in 2008 -- vote by the hundreds and hundreds against the Democratic candidate for President who earns the most popular votes.
So in the event, even a not-so-secret Clinton partisan like Debbie Wasserman Schultz had to go on television repeatedly to say that super-delegates are not earned in caucuses and primaries and therefore cannot be reported by the media in this fashion. In fact, Wasserman Schultz indicated, they shouldn't be tallied at all prior to Philadelphia, at least not as part of electoral-math news reporting.
The Clinton camp said nothing. Misleading reporting about the function and meaning of super-delegates wildly favored Clinton, so as the DNC asked the media to stop what it was doing, the Clinton camp stayed quiet. In fact, they were thrilled that this misleading reporting was -- day in, day out -- making a Sanders nomination seem mathematically impossible. It must have been particularly rich for them to watch John King (CNN) night after election night saying that "Sanders supporters complain" about the reporting of super-delegates -- when in fact it was the very Party apparatus the Clinton camp owed its loyalty to, the DNC, that was doing the complaining.
So all that was fine with the Clinton camp.
Then things got ugly: Sanders starting winning a number of states by large margins, and -- incredibly -- tying or losing the delegate count in those states because state super-delegates refused to support the candidate selected overwhelmingly by voters themselves.
This too was absolutely fine with the Clinton camp.
It was fine with them because it changed the narrative after Sanders' massive victory in, say, New Hampshire -- a state roughly equidistant from Sanders' home in Vermont and Clinton's in New York -- from "Bernie gains massive delegate haul" to "Bernie's win in his 'backyard' nets him zero delegates on Clinton, so who really cares?"
But then something happened. In fact, two things happened.
First, noted pollsters, including Nate Silver of The New York Times, informed the Clinton camp that it was "very possible" she wouldn't be able to win the Democratic nomination via pledged delegates alone. In fact, she might not even come close to the needed 2,383 delegates without heavily relying on Party-selected super-delegates.
Second, the Sanders campaign, seeing the same hard data, concluded -- as a matter of fact, not strategy -- that this meant both Sanders and Clinton would need to make their respective cases to super-delegates in Philadelphia.
It meant that the (apparently convincing) argument Clinton had previously made to super-delegates before any votes had been cast or polls published -- that she had the best chance of winning in November -- would now be made by her opponent, Sanders, and this time with infinitely more data than she had ever had at her own disposal.
Sanders would be able to point to head-to-head state and national polling showing that he had a much better chance than her of defeating Donald Trump in November.
He'd be able to show that he'd won or tied -- or lost so narrowly that another week in-state would have meant a win -- in nearly every November battleground state: Iowa, Illinois, Missouri, Colorado, Minnesota, Michigan, Utah, Maine, Nevada, and Massachusetts. He'd be able to show that Clinton's pre-election, early voting-driven leads had been halved or worse in Ohio, North Carolina, and Arizona -- with Election Day voting in the second of these favoring her only 52% to 48%, and Election Day voting in Arizona actually favoring Senator Sanders, 52% to 48%. She would have to hang her electability argument on Florida -- where she and Sanders perform equally well against Trump in head-to-head polling -- and a bevy of Deep-Southern states the Democrats won't be contesting in November.
So that was when Charles Blow (The New York Times) went on CNN and argued that Bernie Sanders had a "dangerous" view of the role and function of super-delegates.
And that was when other, more ardent Clinton surrogates started spreading the same tale.
Some went so far -- Bakari Sellers of CNN tried particularly hard on this score -- as to call Sanders disloyal to the Democratic Party, implicitly reminding super-delegates that really it was all the money Clinton had raised for their own election campaigns, not anything about her values or first principles, that demanded their loyalty. Sellers and others peddled the tale that somehow Sanders was seeking to destroy the Democratic Party -- even as friends of Sanders like Bill Press (CNN) were telling everyone who would listen that Bernie had always, from day one, said he would not run as an Independent and would support the Democratic nominee for President. No public pledge was required from Bernie -- as was the case with the Republicans -- as in fact Bernie had voted so consistently with the Democrats over the course of his career as an Independent that the idea of doing anything to harm the Party's chance of winning the general election in November was beyond even his capacity to imagine.
So here we are.
Clinton's view of, use of, and benefit from super-delegates has been anti-democratic from the jump.
And now, suddenly, Sanders making a fact-based argument to super-delegates on the very subject they're supposed to care about -- electability -- in the context of an election in which neither party can clinch, is "dangerous."
Fortunately, neither the Sanders campaign nor Sanders supporters appear to be wrong-footed by the audacity of this rhetorical reversal by Clinton, nor by its hypocrisy.
2. The certainty that Sanders would fall way behind in the first half of the nominating season, only to come roaring back in its latter half.
This part of things gets very data-intensive, but I hope you'll bear with me.
First, here's a chart taken from FiveThirtyEight.com data. It shows how polling in the early nominating states -- from the jump -- predicted that Sanders would fall far behind Clinton in the first half of the primary season. Specifically, it ranks, from lowest to highest, Sanders' polling-projected chances of winning the early states, with the actual result of each contest at the head of each subsection.
Figure 1: Sanders' Pre-Election Chances of Winning Past Contests, With Actual Results
(ranked from lowest pre-election chance of winning to highest)
Alabama: 0% chance
Florida: 0% chance
Georgia: 0% chance
Louisiana: 0% chance
Mississippi: 0% chance
North Carolina: 0% chance *
South Carolina: 0% chance
Tennessee: 0% chance
Texas: 0% chance
Virginia: 0% chance
Arkansas: 3% chance
Ohio: 5% chance
Nevada: 29% chance*
* In North Carolina, Election Day voting was a virtual tie (52% to 48% for Clinton). In Nevada, the final margin was relatively small (5.5%). All other Clinton wins were by reasonably healthy margins.
Massachusetts: 6% chance
Illinois: 12% chance
Iowa: 33% chance
Missouri: 46% chance
Michigan: 0% chance
Oklahoma: 49% chance
New Hampshire: 100% chance
Vermont: 100% chance
Note that we begin to see ties in the voting results when Senator Sanders is given even a 6% pre-election chance of winning.
We begin to see Sanders pulling out wins when he's given a 49% or greater chance of winning. The obvious and critical exception here is Michigan, a state Sanders won when he had "no chance" (note that there is no similar result for Clinton to point to).
As the sharp-eyed will note, for one primary, Arizona, there were no pre-election predictions from FiveThirtyEight. And of course, Sanders also won many caucuses in the first half of the nominating season, though none of these were sufficiently polled to result in a pre-election prediction. This of course hurt Sanders, as all of Clinton's wins could be spoken of in advance by the media in flushed excitement, whereas Sanders' wins came in states whose votes necessarily received slightly less media coverage because of the absence of much if any meaningful polling beforehand. Sanders (usually) received ample media coverage when he won states, that is, but he enjoyed only a fraction of the pre-election coverage Clinton did amping up his ability to win in certain states.
The next chart takes the data in the chart above and condenses it -- with the aim of showing that Sanders' chances in the second half of the nominating process are wildly better than was the case in January, February, and March. Again, this data is from FiveThirtyEight (scroll down at link).
Figure 2: Sanders' Odds of Winning, First 21 State Primaries
(with % of contests falling into each category; two largest categories in bold)
No chance (0% chance; 11 states): 52%
Low chance (1% to 29% chance; 5 states): 24%
Reasonable chance: (30% to 70% chance; 3 states): 14%
High chance (71% to 100 chance; 2 states): 10%
So in the first half of the nominating season, polling data gave Sanders little to no chance in 76% of the contests.
That's why it isn't too surprising that Sanders found himself in a deep delegate hole after the first half of the nominating process. But here's the thing: going forward, the Senator has at least a reasonable chance to win -- and in some cases a very high one -- in 89% of contests:
Figure 3: Sanders' Odds of Winning, Remaining 18 State Primaries and Caucuses
(with % of contests falling into each category; two largest categories in bold)
No chance (0%; 0 states): 0%
Low chance (1% to 29%; 2 states): 11%
Reasonable chance (30% to 70%; 13 states): 72%
High chance (71% to 100%; 3 states): 17%
To further refine the data above just a bit, here's a state-by-state breakdown -- with delegate counts -- for how the upcoming races are looking for Sanders, per FiveThirtyEight.com's demographics-based projections:
Figure 4: Breakdown of Sanders' Odds of Winning, Remaining State Primaries and Caucuses
High Odds of Winning (71% to 100% Chance): 53 delegates
North Dakota (90%), 18 delegates
Montana (85%), 21 delegates
Wyoming (80%), 14 delegates
Reasonable Odds of Winning (30% to 70% Chance): 1,484 delegates
Oregon (70%), 61 delegates
West Virginia (67%), 29 delegates
Wisconsin (61%), 86 delegates
Rhode Island (52%), 24 delegates
Connecticut (43%), 55 delegates
Indiana (42%), 83 delegates
New Mexico (42%), 34 delegates
Pennsylvania (41%), 189 delegates
California (37%), 475 delegates
South Dakota (34%), 20 delegates
Kentucky (32%), 55 delegates
New Jersey (32%), 126 delegates
New York (30%), 247 delegates
Low Odds of Winning (0% to 29% Chance): 116 delegates
Delaware (21%), 21 delegates
Maryland (10%), 95 delegates
For those who wonder why I term a 30% to 70% chance of winning a state a "reasonable chance," look at it this way -- if a batter is up at plate in a baseball game and his batting average is .300, would it be fair to say he has a "reasonable" chance of getting a hit? Sure.
So: if Sanders begins getting delegate ties when he has a 6% chance of winning -- a worse chance of winning than he has for any of the remaining 18 state votes -- and, Michigan aside, tends to win states when his chance of winning is 49% or greater (which is the case in 7 of the remaining 18 contests), we can begin to see what the second half of the nominating season could look like for Sanders. Obviously, these data are assurances of nothing; however, they do underscore that, given the polling and demographic data available to both the Clinton and Sanders camps, both sides knew Sanders would storm back into electoral relevance in the second half of the primary campaign.
3. The Clinton camp's use of the media to spin its chances of clinching the nomination via pledged delegates.
Let's be clear: as noted in my last article, Sanders does just fine in states with diverse populations.
But the narrative coming out of the Clinton camp lately -- and therefore the media -- is that because Sanders can only win in a) nearly all-white states, and b) caucuses, he won't continue to perform well over the final 18 contests.
The Clinton camp doesn't believe this, however. Nor does the media.
Not just because it quite evidently isn't true based on the data we have, but because it's an argument whose timing is so wildly convergent with the emergence of serious doubts about Clinton's candidacy that it doesn't so much smack of politics as embody what we Americans hate about all things political.
First and foremost, ten of the 20 whitest states in America have yet to vote, so this is a bizarre time for the media to be saying that Sanders is heading into unfriendly territory demographics-wise.
Somehow, the media has turned Pennsylvania, the 19th whitest state in America, into a state "too diverse" for Sanders to win. Mind you, three of the ten least white states in America were just won by Sanders in landslides, but somehow a state substantially more white than these is a Sanders bogeyman. It's nonsense. The same is true for Wisconsin, which is the 11th-whitest state in America but is being cast as "much more diverse" than the three states Sanders just won in a landslide, when of course it's much less diverse. Why the nonsensical rhetoric from the media? I don't know -- ask the Clinton camp. Most days CNN commentators are (I emphasize, by their own on-air admission) simply reading out Clinton-camp texts and tweets to their viewers. (Dana Bash and Brianna Keilar, I'm looking at you two in particular. Stop reading out campaign propaganda on-air as though it were news -- or reliable.)
Another state that must be shocked to suddenly discover itself "too diverse for Sanders to win" is Connecticut, which is 71% white (the U.S. is 62% white overall). Or Delaware, which is 63% white. While New Jersey and New York are just a hair more diverse than the middle 50% of American states (demographically speaking, as to race and ethnicity) that Sanders is losing to Clinton just 52% to 48% -- New Jersey is 58% white and New York 57% so -- only Maryland looks to pose a real demographic challenge to Sanders, assuming you believe the media canard about Sanders' appeal to nonwhite voters in the first instance. This is probably why Maryland is the state with the lowest "Sanders win" percentage, above.
Demographically, Maryland (54% white) looks much more like the four states Clinton gets 95%+ of her delegate lead from: Florida (55%), Georgia (53%), Mississippi (57%), and Texas (44%). The difference, as Ben Jealous of the NAACP and many others have pointed out, is that nonwhite voters are not monolithic -- the conservative bent of nonwhite voters in the Deep South is not matched by large numbers of nonwhite voters north of the Mason-Dixon line.
In any case, based on past results (see Figure 4 and actual vote tallies from past primaries and caucuses), we can readily imagine Sanders winning states in which he has "high odds of winning" by a margin of, say, 75% to 25% -- which would net him 27 delegates on Clinton's current 230-delegate lead (40 delegates to 13).
While more difficult, it is certainly not beyond the pale to imagine Sanders winning the largest category in Figure 4 -- states in which Sanders has a "reasonable chance of winning" -- by 55% to 45% overall, meaning he would net 148 delegates on Clinton (816 to Clinton's 668).
In the states where Sanders has a low chance of winning -- remembering that, in the very worst of these, Maryland, his chance of winning is still 500% better than was the case in his worst Southern state -- we might expect Sanders to lose to Clinton 55% to 45%, for a net delegate loss of 12 (64 delegates for Clinton, 52 for Sanders).
Under this scenario, Sanders would gain 163 net delegates between now and the end of the California primary (specifically, he'd gain 908 delegates to Clinton's 745).
The result would be that, after California, Sanders would have 1,945 delegates, and Clinton 2,012.
An additional 94 delegates would come from four non-state voting units: Guam, Puerto Rico, the Virgin Islands, and the District of Columbia.
The upshot: the math above emphasizes how easy it will be for Sanders to keep Clinton from clinching the Democratic nomination -- which requires 2,383 delegates -- with just pledged delegates. Indeed, in the very reasonable scenario imagined above, Clinton could win fully 100% of the non-state primaries whose delegates haven't been allotted here and she would still come up hundreds -- literally hundreds -- of delegates shy of clinching the nomination with just pledged delegates.
Or, alternatively, Sanders could slightly outperform the scenario imagined above, win a good portion of the non-state voting, and come out just ahead of Clinton in pledged delegates.
In either scenario, neither candidate would be able to clinch the nomination before Philadelphia.
4. Sanders' growing ability to meet or exceed his delegate targets almost everywhere.
Right now, FiveThirtyEight.com says that Bernie Sanders has met 92% of his delegate target -- the number he needs to ensure that he wins the pledged delegate battle -- after braving far and away the most difficult portion (for him) of the primary calendar. Meanwhile, Clinton is only 8% over her target, despite having had the benefit of all of the states that are most favorable to her having already voted.
Here another view of this:
Figure 5: Sanders' Performance, Relative to Delegate Targets, in States Voting Before March 5
(with performance relative to target, followed by % of all pre-3/5 states falling into the given category; largest category in bold)
Exceeded Target: 19%
Met Target: 6%
Failed to Meet Target: 75%
Figure 6: Sanders' Performance, Relative to Delegate Targets, in States Voting On or After March 5
(with performance relative to target, followed by % of all pre-3/5 states falling into the given category; largest category in bold)
Exceeded Target: 47%
Met Target: 16%
Failed to Meet Target: 37%
When you go inside these numbers, they're even more striking.
Of the 16 contests before March 5, Sanders only exceeded his delegate targets three times -- but even in these three successful contests, he never exceeded his delegate target by more than two delegates (out of a total of 4,051 pledged delegates to be awarded this primary season). So even when Sanders did well, he didn't do so in dramatic fashion.
From March 5 onward, there have been 19 contests. Sanders exceeded his target in nine of them -- and in every case but one, exceeded his target by more than two delegates. In many cases, much more.
So really, it isn't enough to say that Sanders is doing better now than he was, but that the media expected him to be doing better now than before and -- even still -- he's outperforming even those out-sized expectations. FiveThirtyEight.com predicted, for instance, that Sanders would win both Hawaii and Alaska by 8 points, and instead he won the former by 40 points and the latter by 60. Essentially, the media hasn't been able to keep pace with Sanders' momentum, even in those rare instances it uses hard-data projections to try.
One consequence of this is that many of the projections for future Sanders delegate hauls are almost certainly wrong. First, they're wrong in caucus states because those projections (e.g., from FiveThirtyEight) still project 55%-45% Sanders wins in these votes, even though Sanders has already shown that his caucus wins tend to be by 40 to 50 points rather than ten. Second, the future projections are wrong because they do not, in many instances, consider past voting behaviors; for instance, whereas demographic data shows that Sanders has a 41% of winning Pennsylvania, for some reason FiveThirtyEight.com has used just a single poll from the last four weeks -- which poll sampled just 408 likely voters in a state of 13 million -- to conclude that Sanders actually only has a 4% chance of winning.
Having only one poll from the last month with which to predict the results in Pennsylvania means using, instead, data that comes primarily from the first half of the election cycle -- when all factors and data seemed to favor Clinton -- to predict an election that's happening under radically different circumstances.
So let's repeat here what I said at the start of this article: Sanders will win more pledged delegates in the second half of this nominating season than Hillary Clinton. In fact, he'll almost certainly win more pledged delegates than Clinton in the final three and a half months of the primary season. He'll also close out the primary season, it appears, beating Donald Trump by substantially more than Clinton in nearly every national and battle-ground state poll taken.
So pollsters now have a choice: use polls from a month or more ago to predict the future -- which FiveThirtyEight is now largely doing, and which the Clinton camp wishes for them to continue doing -- or use the results as they've actually come in from individual states, not just the "light-blue versus dark-blue" binaries pushed by CNN's simplistic "magic wall" but the final tallies everywhere, to correctly chart the course of an historic electoral comeback.
It's a narrative the Clinton campaign will be pushing back on almost every single day, usually with audacious claims that are offensive to anyone with eyes and ears -- such as the recent one to the effect that Bernie Sanders is running a "very negative" campaign. If the Clinton campaign is so stridently disrespecting voters' intelligence in March, how will they distinguish themselves later on against a man who's made his political career disrespecting voters' intelligence, Donald Trump?
The simple answer, I suppose, is that if they keep on the way they are, they'll never get the chance to find out.
Seth Abramson is the Series Editor for Best American Experimental Writing (Wesleyan University) and the author, most recently, of DATA (BlazeVOX, 2016).