For the first time since February, there is only one state holding a primary during a week. It’s also one of the biggest states in the country — New York. It also comes at a surprisingly crucial time during the campaign. By mid-April, the norm is that the race is over. The lack of money for trailing candidates has typically forced them to suspend their campaign and party leaders are pressing for unity behind the likely nominee. This year, the race is different. Bernie Sanders has enough money to keep running through the convention. Republican leaders are definitely not pushing for unity behind Donald Trump.
On the Republican side, we have seen the rules that New York is using in earlier states. Delegates are awarded “proportionately” by congressional district (three in each of the twenty-seven districts or eighty-one total) and statewide (fourteen delegates). As in many states, it takes twenty percent to become eligible for delegates, and a district (or the statewide results) becomes winner-take-all at fifty percent. As in many states, the congressional district is a 2-1 split between first and second place if two or more candidates qualify. At the state level, the party rounds delegates to the nearest whole number. If there are any delegates remaining, they go to the winner. If there are too many delegates allocated, the additional delegates will be taken from the last-placed candidate. (At most, the math should lead to one or two delegates being added or subtracted.)
Heading into the primary, Donald Trump seems to be flirting with fifty percent state-wide. By mathematical necessity, if he gets over fifty percent state-wide, he will get over fifty percent in some districts. Additionally, Ted Cruz has the small problem of having attacked “New York” values while he was running in other states. He can probably convince upstate voters (and how you define upstate depends upon where in the state you live — for New York City, upstate includes Westchester and Rockland County, but for Albany and Syracuse voters, Westchester and Rockland County are part of the New York City area as is Long Island) that he meant New York City, not New York State. But only nine districts are wholly upstate (by the narrow definition). Perhaps, he can convince some New York City Republicans that he meant the values espoused by Democratic politicians, but Cruz is not likely to be competitive in the New York City districts.
Given this background, the issue is how the campaigns focus their resources for the last two weeks. Trump has succeeded (to the extent that he has succeeded) by blunt force with very little use of internal polling to narrow his focus. Cruz and John Kasich are probably going to rely heavily on internals — pounding Trump in the districts where Trump is close to fifty percent. In districts, where it is clear who will take second place, the third-placed candidate will probably focus most of their money on attacking Trump. It is only in the districts where there is a battle for second that you should see Cruz and Kasich attacking each other. Barring Trump doing poorly in upstate New York, he will probably get close to or over seventy delegates out of New York. If Trump gets over eighty delegates, he will have mostly made up for his poor performance in Wisconsin and Colorado (and his likely poor performance in Wyoming). If Trump falls to second in some districts (with the result that he falls under sixty delegates total), it is hard to see how Trump gets to striking range of 1237 and the odds of a contested convention will increase.
On the Democratic side, it is the normal proportional rules. The delegate per congressional district range from five to seven delegates (one hundred sixty three total). In the four districts (11th, 19th, 22nd, and 23rd) with five delegates, a simple majority gives a 3-2 split, but it takes 70% of the vote to get a 4-1 split. In the five districts (3rd, 7th, 14th, 20th, and 26th) with seven delegates, a simple majority gets a 4-3 split, but it takes 64.3% to get a 5-2 split, and a 78.6% to get a 6-1 split. In the remaining eighteen districts with six delegates each, it takes 58.35% to get a 4-2 split, and 75% to get a 5-1 split. As the recent results in Wisconsin and Wyoming showed, it is hard to pick extra delegates in the congressional districts, particularly the “even” districts. In those two states, Sanders only picked up a net of five delegates at the congressional district level (with three of those delegates coming from winning the three “odd” districts in Wisconsin and a pickup of one extra delegate — creating a net of two — from the margin in one district — out of nine districts total).
Unlike many of the smaller states, however, New York is large enough that the state-wide result could have a big impact on the total delegate count. New York has thirty pledged party leader delegates. It only takes 51.67% to get a 16-14 majority, and a candidate gains an additional delegate for every 3.34%. New York also has fifty-four at-large delegates. It only take 51% to get a 28-26 majority, and a candidate gains an additional delegate for every 1.9%. Thus a 10% margin in New York creates a 17-13 split in party leader delegates and a 30-24 split in the at-large delegates (and more likely than not winning most of the odd districts getting an 86-77 split at the district level for a total net of 19 delegates). Given that Sanders is currently over 200 pledged delegates behind Hillary Clinton and that New York represents slightly over one-seventh of the remaining pledged delegates, New York is key for both candidates. A Clinton win (particularly if it is a double digit win as shown in the current polls) will make it even more difficult for Sanders to have any chance at catching up in the pledged delegate count. A double digit win for Sanders would give some credibility to his argument that he might be able to get close enough to put some pressure on the unpledged delegates.
The good news in the math for Sanders is that there are only about 200 unpledged delegates who have yet to endorse a candidate. With so few available, it will be some time before the remaining unpledged delegates could bring this race to a close. Over the next two weeks, there are only 600 pledged delegates at stake. Even if (as current polls suggest), Clinton does well in these states, she is probably only looking at 330-350 delegates. The May primaries involve approximately 240 delegates, but these states are probably better states for Sanders than the states over the next two weeks. However, after the last of these states vote on May 17, there will only be 800 delegates left. If Clinton has over 1700 pledged delegates and is leading by over 200 pledged delegates, it will be almost impossible for Sanders to close the gap with the June states (as he would need to take over 60% of the delegates). At that point, the remaining unpledged delegates might decide to intervene to bring the race to an end before the June 7 primary. In short, even though Sanders might be able to close the gap somewhat in May, he needs to keep Clinton from increasing the margin over the next two weeks if he wants to keep the race going through June 7.