Sanders Will Likely Win Through June

The story in the newspapers is that Sanders’ West Virginia victory means that the Democratic race will extend until July.  This was always going to be true, as the Sanders camp has said since New York, regardless of the outcome in West Virginia.

The demographics in May 17 races suggest big wins for Sanders in Kentucky (similar to WV) and Oregon (similar to Washington, where he won in a landslide).  Those wins should net him nearly 45 delegates, and he may come close to the 70+ percent of the delegates that he needs to stay on pace to tie.

Wins will help the public narrative on Sanders’ viability, as well as his chances on June 7 primaries in California, Montana, North Dakota, New Jersey, New Mexico, and South Dakota.  Unfortunately for Sanders, polling in California and New Jersey shows Clinton with sizable leads there, and we project her heading into the convention with a 350-delegate lead and needing around 180 superdelegates to clinch the nomination.

Sanders will likely win through May, but polling in California and New Jersey portends the end for Sanders.
Sanders will likely win through May, but polling in California and New Jersey portends the end for Sanders.

None of this, of course, speaks to core Sanders supporters. Running a projection in which Sanders wins 100 percent of all votes in states where there are no state-level polls (i.e. all but New Jersey and California), we still show her heading into the convention with a 180-delegate lead.

If Sanders wins 100 percent of all votes in states where there are no polls, he still comes up short.
If Sanders wins 100 percent of all votes in states where there are no polls, he still comes up short.

Sanders will not be mathematically eliminated until June 8,  and so does retain a chance of taking more delegates than Clinton into the convention.  To do so, he needs to perform better than he did in Indiana and West Virginia, and will need Oregon-sized wins throughout.  That would represent a nearly 50-point swing in California from current polling, and a 60-point swing in New Jersey.  Winning 60% of the vote in those states and 70% in the remaining still leaves him 70-delegates short.

Bottom line:  It’s not overly dismissive to say that Clinton will be the Democratic nominee, but the news this week and next will suggest a closer race.

Acela Primaries Preview: Trump sweep, Clinton ahead (just)

There was news this week: a possible Cruz-Kasich coordination, the possibility Sanders would shift after today’s primaries, etc. None affected the delegate math, hence the quiet week here.  After tonight’s results, we’ll put in state-by-state benchmarks into the model to give a better preview of where we expect the races to be at each state.  Tonight looks fairly straightforward on the Republican side, with the possibility for some surprises on the Democratic side.

Republican race

Trump should win each of the states up for grabs, approaching 50 percent of the vote in each of them.  Wildcards will be how individual high-income CDs in Connecticut vote, but overall we expect 90 of the 117 delegates up for grab tonight.

Trump looks for an April 26 sweep, keeping him close to on track for 1,237.
Trump looks for an April 26 sweep, keeping him close to on track for 1,237.

This performance keeps Trump close to hitting the magic 1,237 mark.  Pennsylvania has 54 unbound delegates up for grabs, and a strong showing there can help Trump make the case that they should go his way, keeping him just on track to win on the first ballot.  Much depends on Indiana, and, of course, California.

Democratic Race

After the strong showing in New York, the narrative has been that Clinton would effectively finish Sanders off during the Acela primaries.  Don’t count on it. Our model has her losing Rhode Island, and Delaware and Connecticut are too close to call.

While her more-comfortable leads in the more-important Maryland and Pennsylvania races should give her around a 30-delegate pickup on the night (Clinton should win around 208 to Sanders’ 177), the media narrative has been that she will sweep. Failure to do so may sustain the current competitive narrative.

Sanders may win three states on April 26, but Clinton will likely win the delegate count.
Sanders may win three states on April 26, but Clinton will likely win the delegate count.

At the end of the night, we expect Clinton to have 1,660 pledged delegates to Sanders’ 1,376. There are still 1,016 delegates at stake following tonight, but Sanders would need nearly 64% of them to tie Clinton before the convention. That’s an increase from the 60% he needs before these primaries, and we don’t expect him to hit that mark in any of the five contests.

New York: Big Trump, Clinton wins don’t change Delegate Math

Both Hillary Clinton and Donald Trump had big nights in New York, both getting around 60% of the vote.  They both needed the big wins, and they got them.  Clinton won 139 of 247 delegates, and Trump will get around 90 of the 95 on the GOP side.

They didn’t, however, deviate too much from our projections (although, and we admit it, we went on the conservative side on both, the mid-point part of our model would have had us closer, at Clinton 105 and Trump 85. Oops.)

Republican Race

With Trump’s big win, we now estimate he will get 1,160 of the 1,237 delegates he needs to lock up the nomination on the first ballot.

New York handed Trump his biggest step up since March 15, but he still looks to finish short.
New York handed Trump his biggest step up since March 15, but he still looks to finish short.

There are a lot of caveats to this analysis. Most importantly, there’s vanishing little polling in the upcoming states.  While he should get 40% or so of the votes in Maryland, Connecticut, Pennsylvania, Delaware, and Rhode Island on April 26, the big outstanding tests are Indiana and California. Our model currently has Trump losing New Jersey.  This is almost certainly wrong, and would hand him 51 delegates (or 1,211 of the 1,237 he needs), but is based on now 2-month-old polling.

Another caveat: one important outcome for the evening is that Ted Cruz held steady with 546 pledged delegates, by our count. With just 618 left on the board, he joins Kasich in becoming mathematically eliminated.  However, Kasich’s reasonably strong performance in New York shows that’s not necessarily a killer, but our model has Cruz cleaning up in states like Montana (and winning Indiana), which could be more difficult if he’s having to wage a Kasich-like battle for relevancy.

Democratic Race

With California at the end (well, DC) of the calendar and 714 super delegates able to change allegiances through the convention, Sanders will never face mathematical elimination. New York, as expected, made his case much more difficult.

Clinton's win in New York all but assures her of the nomination, but the race stays tight through the convention, where she'll need superdelegates.
Clinton’s win in New York all but assures her of the nomination, but the race stays tight through the convention, where she’ll need superdelegates.

Like Trump, Clinton now enters a slate of states that should be favorable.  Our model has her picking up between 20-30 delegates over Sanders in next week’s primaries, which look a lot like New York in terms of polling (10-20 in Maryland and 10 in Pennsylvania).  It’s not enough to clinch the nomination with pledged delegates alone.

The Democratic race should stay close, but Sanders should remain around 300 delegates behind Clinton.
The Democratic race should stay close, but Sanders should remain around 300 delegates behind Clinton.

For Sanders, we show him getting around 48% of the remaining delegates.  He needs 59% of the 1,400 remaining delegates to work this to a tie.

His campaign is sending mixed messages on his way forward.  Sanders’ campaign manager, Jeff Weaver, told MSNBC that Sanders would “absolutely” try to flip superdelegates if he remained behind in the popular vote and delegate count.  They would need to win 500+ of the 714 superdelegates to do so.

At the same time, campaign strategist Tad Devine left open an exit, telling the AP that the campaign planned to “sit back and assess” the campaign’s chances after next week’s contests.



New York Primary Baseline: Clinton 130, Trump 80

New Yorkers vote today in what the press is selling an electoral pivot.  We think next week’s elections are bigger in terms of the delegate math, and there’s sufficient (and pretty good) polling predicting what will happen today.

Republican Primary

Our model predicts that Donald Trump will get 52.1 percent of the vote, John Kasich 26.7 percent, and Ted Cruz 21.2 percent. New York is winner-take-all if a candidate hits 50% of the vote, so Trump’s final number matters a lot.

Why, then, do we have him at 80 delegates? Well, only 14 delegates are assigned given the statewide vote.  The remaining 81 are allocated by races in the 27 individual Congressional district races.  Again, win 50% of that vote, and you collect all three delegates.  If you win with less than 50%, you get 2 delegates and your closer competitor gets 1.

Only Optimus appears to have strong CD-level polling.  Their poll from last week shows Trump winning ALL of the CDs, but missing the 50% threshold in 14 districts.  In one district, NY-12, Trump polls around even with John Kasich.  Put this together, and 80 delegates appears to be the most-conservative estimate (Kasich is on pace to pick up 12 of those that Trump misses, according to the same poll).

Our blind spots are legion. Our model incorporated Kasich doing particularly well in the 14 CDs that voted a Democratic house member in in 2014 by more than 15 points, but the polling data doesn’t match this 1 for 1. While Trump rarely exceeds his poll numbers, in New York it may be easier for him to do so, which could push him over the edge.  We’ll know more at night’s end, but for now, it’s safe to say anything under 80 delegates is a bad night for Trump, and it’s still possible for him to get all 95.

Democratic Primary

Our model shows Hillary Clinton winning 54.6% to Bernie Sanders’ 45.4%.  This tracks with most recent polls, which show Clinton with a 10- to 18-point lead.

Running that prediction straight would give Clinton a 135-112 delegate haul.  However, only 84 of the 247 delegates will be allocated based on the state-wide vote.  The remaining delegates are elected based on the proportion of Congressional districts each candidate wins.  18 of the 27 districts have 6 delegates each up for grabs, 5 CDs have 7 delegates, and the remaining 4 CDs have just 5.

For the districts with an odd number of delegates, if the candidate wins 50% + 1 vote, they collect the “bonus” delegate.  For the 18 CDs with 6 delegates, however, the winner needs to get 58.34% of the vote in that district to take 4 delegates; anything below that threshold will result in a 3-3 tie.

What does this mean for the delegate math? Should Clinton win statewide, and in each CD, by a 56-44 margin, that would be an absolutely huge electoral victory and news story, but would net her only 19 delegates over Sanders because they will have split 18 CDs 3-3.  However, should she poll 58.5-41.5 (this is in no way likely), she would net 46 delegates over Sanders, winning an extra delegate in all but 4 CDs.

Back to the model:  we show Clinton picking up 46 of the 84 statewide delegates (8 net).  We’ll assume a tie in each of the 18 CDs (0 net), and that she picks up a net delegate in the 9 “odd” CDs, to net a total of 17 delegates over Sanders.  This would result in a 132-115 delegate result.  Since this is a “good night/bad night” calculation, we’re flipping one CD to a Bernie blowout, given the huge crowds he’s collected in the state and to account for missed polls.  If Clinton pulls in 130 delegates or more, it’s a good night.  If she gets fewer, she had a bad night.

If she loses the state, the race might not be over next week.

Clinton Likely To Net 25 in New York

Ten polls in the past two weeks show Hillary Clinton maintaining a double-digit lead over Bernie Sanders in New York.  With our model set to Sanders winning 80% of the undecided vote, we predict Clinton will win the New York primary with around 55 percent of the vote.

It will likely take weeks to get the final delegate count, in part because New York is proportional by Congressional district, and the district maps in the city cross all kinds of political boundaries. In CDs that have six delegates available (18 of 27), a candidate needs to exceed 58% of the vote to avoid splitting the delegates 3-3.  This likely helps Sanders by around 10 delegates statewide.

Since we don’t have CD-level polling, and our model doesn’t account for demographics in any meaningful way, it’s not possible to drill down into how each of the 27 CDs may split.  Nevertheless, our estimate shows Clinton netting roughly 25 delegates in the night, which would extend her pledged-delegate lead over Sanders to around 240. The win is what will set the narrative for the remaining races, but a 20+ Clinton delegate victory will come close to sealing the deal.

Our model shows Clinton netting an addition 42 delegates in the April 26 contests, mostly on the strength of an 18-point win in Maryland and a 6-point win in Pennsylvania, based on recent polls.  Even were she just to net 30, she would have a 270-delegate lead.  While not mathematically impossible to overcome until June 7, Sanders would need to post 50-point victories in each of the remaining contests to work the race to a tie.  He will likely do so in Oregon, but in California, our model predicts a 4-point Clinton win, meaning Sanders has a lot of ground to make up in the next two months.

The “good” news for Sanders, though, is that the polls also show him keeping Clinton around 200 delegates short of what she needs to win the nomination on pledged delegates alone.  It’s very likely that she has the super delegates needed to carry her over on the first ballot, but without Clinton hitting the 2,383 magic number in the primaries, Sanders can legitimately fight through the convention.


Sanders’ Magic Number: 70

We were going to start doing a state-by-state countdown to make it clearer when it was mathematically impossible for Donald Trump to clear the 1,237 delegates he needs to clinch or where Bernie Sanders might overtake Hillary Clinton in pledged delegates.  Turns out, Ted Cruz is right: “California will decide“.  There’s simply too many delegates there; as unlikely as it might be, if Trump or Sanders win 100% of the vote there, they win.

A lot of folks are talking about how Sanders needs to win 57% or more of the remaining delegates to tie. That number doesn’t incorporate the polling that shows her with 10-point leads in New York, Maryland, Pennsylvania, and California (there is polling for New Jersey, but it’s really old).

Using a very simple model that assumes a Clinton “ceiling” (i.e. all undecideds go to Bernie, so if Clinton is polling under 50 percent as she is in California, Sanders actually wins) using the current poll averages, Sanders would need to pull in 70 percent of the vote in all other jurisdictions to pull ahead of Clinton by June 14.

Based on past results, that seems doable in Oregon, but unlikely in, say, Puerto Rico (which has basically the same number of delegates).

Later this week, we’ll add calculations to show how much of each vote the candidates need in each state to hit their targets, but with the states with polling representing around 70 percent of the outstanding delegates, Sanders must lower Clinton’s ceiling substantially in New York, Maryland, and Pennsylvania to have a realistic shot of catching her.

Wisconsin primary results: Trump, Clinton less likely to clinch

Sanders exceeded our pre-primary delegate estimate, picking up 48 instead of the 47 we projected to net a total of 10 delegates.  Cruz also exceeded expectations, hauling in 36 delegates instead of the 30 we projected.  The error there was all Kasich underperforming; our estimate showed him performing stronger in the 2nd and 3rd Congressional districts, where it look like he will have actually come in third.

The results don’t alter our convention projections much, which point to a likely contested convention for the Republicans.  On the Democratic side, it remains unlikely that Sanders will catch Clinton in pledged delegates, but it does look like superdelegates will have to get the winning candidate over the nomination threshold:

Sanders win cuts Clinton's lead to around 250. She has a chance on April 19th and 26th to extend the lead in such a way that make it nearly impossible for Sanders to catch her.
Sanders win cuts Clinton’s lead to around 250. She has a chance on April 19th and 26th to extend the lead in such a way that make it nearly impossible for Sanders to catch her.
GOP race post Wisconsin
Trump’s growth is stunted, but a winner-take-all for New York’s 95 delegates could have him on his way.

Sanders will likely do well in the small, western caucus state of Wyoming this weekend, but with only 14 delegates in play (and notable differences between the electorate in Wyoming and New York), it’s hard to see that contest changing the race considerably.

Instead, all eyes turn to New York.  On the Democratic side, our model projects a 10-point Clinton victory to net 23 delegates (winning 135 to Sanders’ 112).  State-level polls from Maryland and Pennsylvania also show her leading in those big April 26 states where we project her to pick up another 28 delegates.  Those two weeks would erase Sanders’ impressive rally over the past few weeks.  With Sanders’ big Wisconsin win and the age of the data in the upcoming states, it’s easy to see those numbers changing – and with them the completion of the race.  New York is, in essence, Sanders’ last shot to change the race.

On the Republican side, our baseline projection has Trump, who’s polling at 51.4%, picking up all 95 delegates.  New York awards 81 of its delegates at the Congressional district level, and the remaining for the statewide winner.  In 2014, 14 of the 27 Congressional districts elected a Democrat to the House of Representatives by more than a 10-point margin.  It’s difficult to see those districts giving over half of their vote to Trump to trigger the winner-take-all rules there, so it’s likely that Cruz (or Kasich, about whom we’ll have to write a more-speculative column) chip into the 95-delegate haul.  Our most-generous projection predicts Trump getting 1,192 of the 1,237 delegates he needs before the convention, with 122 delegates unbound on the first ballot.  A 44-delegate margin may put him in striking distance, but losing delegates on some blue districts in New York (and, in June, in California) could put him just out of reach.

Wisconsin projections

A quick word on what to expect tonight:  Cruz is likely to win Wisconsin, and may do so by nearly double digits.  Our delegate projection shows him wining around 43% of the vote, with a delegate breakdown of Cruz 30, Trump 9, and Kasich 6.  Trump and Kasich will depend on winning Congressional districts.  While Kasich will likely do well in blue districts near Milwaukee and Madison, it’s isn’t unreasonable to expect Trump to underperform the projection above if he’s lagging behind Cruz so far in the statewide tally.  This projection has trump ending up short of the 1,237 delegates he needs by around 40.  The key to watch for will be whether Trump picks up fewer than 9 delegates.  If he does, expect “Trump has lost momentum” stories to increase through New York.

On the Democratic side, our projection shows Sanders winning by 10 points, taking 47 delegates to Clinton’s 39 to chip away another 8 delegates into her 200-delegate lead.  That, however, won’t be the story if he wins by such a margin.  The story will be whether Clinton can hold on to the 10-point margin she currently enjoys in New York.  If she does, she’ll net 23 delegates and mostly erase Sanders’ gains over the past few weeks and effectively cut off any chance he has of catching her in the pledged delegate count.  Sanders knows this, and is investing heavily in both Wisconsin and New York to keep his campaign alive.  The key to this race will be whether Sanders captures more than 55 percent of the vote.  Doing so will give him a decent delegate margin in the state, and probably signal more difficulty for Clinton in New York.


Sanders needs a 30-point swing in New York

A lot has been written over the past week about Bernie Sanders’ chances of getting a majority of unplugged delegates.  Nate Sanders has a good writeup, but doesn’t show a lot of his work. would be a better tool if it defaulted to current polling.  Here’s our contribution to this theme:

Bernie Sanders' unlikely path to pledged delegate majority.
Bernie Sanders’ unlikely path to pledged delegate majority.

A few assumptions at the top:

  • Our model has Sanders winning Wisconsin 54.6-45.4, giving him +8 delegates.  We’re keeping that assumption.
  • Few states have state-level polling (though the states that do account for 70% of outstanding delegates).  Our model uses national polls there, which would give Clinton a 53.2-46.8 edge.  We’ll ignore that in every state, with most states going 60-40 for Sanders, which means he needs to over perform his polls by 15 points.
  • For a few states (Wyoming, Oregon, Montana, New Mexico, DC), we’ve assigned Sanders Alaska-level blowouts.
  • We’re putting Clinton’s delegate lead at 229 through April 1.

Our rough estimate is that Sanders would need to take 15 points off of Clinton in New York to have a chance of breaking even, even with the generous spot our assumptions give him.  The latest polling average has Clinton up there 57-37.  For this exercise, we have Sanders winning 58-42.  Here’s how this scenario plays out, and look closely at what a polling shift would need to happen to get there:

An incredible change of fortune for Sanders gets him an edge over Clinton, but both would need super delegates to decide.
An incredible change of fortune for Sanders gets him an edge over Clinton, but both would need super delegates to decide.

We left Maryland and Pennsylvania within the reasonable bounds of change, and assigned Sanders a 10-point victory in California (our model shows Clinton actually just squeaking by there by one point).

Here’s a more-likely scenario:

A more-realistic Bernie surge leaves him about 90 delegates short.
A more-realistic Bernie surge leaves him about 90 delegates short.

For this chart, we leave in place most of the favorable projections from the last run, but instead leave Clinton winning New York 52-48 and giving her the 1-point victory in California, a conservative estimate on the latest polls. This still represents Sanders taking away 5 points of Clinton’s support in New York, which a 10-point win in Wisconsin might reasonably do (despite Clinton’s heavy campaigning in the Empire State).  The final tally in this scenario would be Clinton 2,061 delegates, Sanders 1,979.

The key here is that each percent of the vote in New York is worth about 2.5 delegates.  If Clinton nets more than about 10 delegates there, it’s very difficult to see how Sanders can break even.

Trump needs New York, California, and Unbound Delegates

Our model takes into account the state-specific delegate selection rules and the state-specific poll averages (if they’re available).  In roughly half of the plausible scenarios we run, Donald Trump comes up around 100 delegates short of the 1,237 he needs to win on the first ballot. In real life, this probably requires more strategy than #NeverTrump actors have heretofore realized, but makes Cook Political/FiveThirtyEight’s “two-front war” look like what the electorate is already planning.

If Trump fails to exceed his poll numbers and Kasich wins districts in which a Democratic representative won by more than 10 points in 2014, Trump gets 1,147 pledged delegates.
If Trump fails to exceed his poll numbers and Kasich wins districts in which a Democratic representative won by more than 10 points in 2014, Trump gets 1,147 pledged delegates.

The chart above is a reasonably likely outcome:  Trump fails to exceed his poll averages and John Kasich wins congressional districts in which the House member won by over 10 percentage points in 2014.


Above is the math behind that projection.  A couple of notes:  Colorado is just a guess for delegate outcomes, the state will be a black box.  The model has Trump winning 100% of the Congressional districts, in New York’s district winner-take-all contests based on the strength of his state-wide polling (he’s a good 35 points ahead of Kasich), but it’s hard to imagine Ted Cruz and Kasich failing to pick up. That’s probably balanced by Katich’s stronger showing in California, where the model shows him picking up 3 delegates in 30 of the 53 House districts.

Closer examination shows a big gap in Pennsylvania’s delegates.  That’s because voters select unbound delegates on the district basis.  West Virginia follows a somewhat similar model, where delegates may or may not declare their intention on the ballot.  There are already 75 unbound delegates heading to Cleveland, and another 174 bound to candidates who have “suspended” their campaigns.  That’s a big wildcard:  enough to push Trump over the edge if a majority break for him.

This outcome actually requires Kasich to stick around and remain viable through June.  Kasich can’t win the 1,237 needed on the first ballot; there just aren’t enough points left on the board, so a vote for Kasich is just a “none of the above” vote.  Sure, he beat Trump in Utah, but he lost to Rubio – who had dropped out a week earlier – in Arizona due to the prevalence of early voting.  It isn’t too big of a stretch to see the “mathematically unviable” argument swaying enough voters in Pennsylvania, New York, and California to make his spoiler status not that big of a deal.

It’s also not a sure thing that this is the direction late-deciders break.  The  models above show Trump picking up 0% of undecided voters.  If Cruz picks up 60% and Trump picks up just 20%, the race looks a lot different:


In that projection, Trump falls just 3 delegates short, and it’s easy to imagine him picking those up among the unbound delegates.

If, based on that slightly weaker projection for Kasich, he loses more of the Democrat-heavy congressional districts in California, with Trump and Cruz picking up a proportional share, Trump wins outright:

Cruz Momentum with no Kasich CD bonus

None of this takes into account the political realities that govern the later elections and a brokered convention itself.  As Trump has picked up some endorsements (and Cruz only some half-hearted ones), GOP leaders could fear the “riots in the streets” outcome, or fear that by suppressing Trump the party could permanently lose the voters he’s won. It also assumes that almost no delegates jump on the Trump bandwagon in subsequent ballots.  Nevertheless, it shows there’s still a reasonable set of outcomes in which Trump fails to secure 1,237 delegates before Cleveland, if the GOP plays its cards right.