Nate Silver again has failed to do the delegate math on the GOP primary.
Trump needs to get from his 25-35% base to 51% at some point. Large % of GOP voters have ruled him out. Not many “swing voters” left. (3/)
— Nate Silver (@NateSilver538) February 18, 2016
As I have already pointed out, this is mathematically incorrect. The GOP primary rules this year are setup so that a candidate can easily win a majority of the delegates with only a plurality of the vote. Due to arcana of the rules, 20% is the magic threshold to pass. 25% is safer. 35% – which Trump’s national polling average has hovered around for the last month – is a lock.
This does, however, have one rather strict requirement. Winning the GOP nomination on a plurality requires at least a 3-man race. If any of Trump’s opponents can whittle it down to a one-on-one race, they have a chance. As Mr. Silver has already noted, when candidates drop out has a huge influence. The other important influence is what candidates drop out.
The current GOP field has already narrowed considerably. However, it still consists of six candidates. In this environment, Trump’s 35% plurality continues to make him a shoe-in for the nomination. So let’s consider each of the remaining candidates and what happens if and when they drop out.
First, let’s take a look at the current national RCP average. Yes, I’m aware that the state level polls matter far more. But we can use this as a current proxy for our thought experiment.
- Trump – 33.8%
- Cruz – 21.0%
- Rubio – 16.3%
- Kasich – 8.8%
- Carson – 6.0%
- Bush – 4.5%
First note that the polling average only sums to 90.4%. That leaves 9.6% of the electorate undecided. Mr. Silver makes the assumption that Trump won’t do particularly well with these undecideds. I haven’t seen him explicitly state it, but I would guess that he assumes Trump will underperform his current poll numbers within that group. He may in fact be right. However, I don’t think there’s any good reason to assume that as of yet. For the sake of our experiment, we’ll forget about those voters and just work with the numbers that we have.
Sooner or later most of the candidates still left will drop out. Even with six candidates in the field, single digit poll numbers won’t win you the nomination. Remember: you can’t even be entered for nomination at the convention unless you’ve outright won at least eight states, and single digit numbers won’t do it for you. “Momentum” strategies can work, but three out of the five remaining candidates are relying on a momentum strategy. It will only work for one of them.
The South Carolina primary is in two days. None of the remaining candidates will drop out before then. If they were going to do so, they would have already. Will any drop out afterward? The only dropout case that I see potential for is Bush. Right now the South Carolina RCP average shows him 6 points below Rubio and only one point (a statistical tie) above Kasich. His national numbers are likewise abysmal, and now there are rumors that his campaign is running out of cash. At the same time, Rubio is lining up the endorsements that Bush was counting on.
Logically, if he finishes behind Rubio – and especially if he finishes behind Kasich – Bush should bow out after South Carolina. My prediction? He won’t – unless he really is out of money. Everything to come out of his campaign to date shows that he – or his family; it’s unclear – really and truly just doesn’t understand that the current crop of GOP voters simply does not want him. He’ll push for one more (Nevada) if it’s possible for him to do so. Until his ego finally lets go, he’ll continue to split the “establishment lane” vote.
Likewise, Kasich won’t drop out yet either. He’s still riding a bounce from a) overperforming in New Hampshire and b) being one of the few men left standing. In a few short weeks he’s gone from “who?” to… well, doing better than Bush. My guess is that he’s going to try really hard to hold on until March 15th when his home state of Ohio votes. He’ll be hoping that “favorite son” status will help him pull out an actual win. It’s very questionable, though, if he’ll have the funds to hold on. My guess? He’ll either be out on the 15th or he’ll be out sooner after a dismal performance on Super Tuesday (March 1st). Until then, he also continues to split the establishment lane vote.
Rubio is in until Super Tuesday no matter what. He’s got a strong Iowa performance to point to and all signs point to him performing well in South Carolina. There’s no way he’ll catch up to Trump there, barring a catastrophe. But he’s in a statistical dead heat with Cruz for second place and he very well might pull that off with the South Carolina establishment behind him. Even if he doesn’t, he’s likely to be a very close third. The potential upset? If Bush somehow manages to pull out the upset and close the 6 point gap for second or even third. Even then, though, expect Rubio to hold on until… you guessed it, Super Tuesday. It’s only a week and a half away and Rubio’s got cash. He won’t be out before that. He’s also unlikely to be out before March 15th (that date again!) when his home state of Florida votes. Again, it’s not that far off and he’s got the cash. If he does well between now and then, he’ll probably be in for a while at that point.
Cruz looks poised for a 2nd place finish on Saturday in South Carolina. If things go poorly for him, that may slip to third. I don’t expect it. He proved in Iowa that he knows how to run a killer ground game, and word is that he’s brought his A game to South Carolina. I expect him to overperform his polls – but not by nearly enough to catch Trump’s 16 point lead. But even a disappointing third place finish keeps him in the race as the only non-Trump candidate who’s actually won a state. Plus, he’s got plenty of cash and he’s proven to be disciplined and strategic at spending it. Unless his poll numbers tank, Cruz is likely in it through March.
That leaves Carson. Honestly, who knows what Carson is thinking right now? However, he’s now vowed to continue to Nevada no matter what happens in South Carolina. And Super Tuesday is only three days later. So I would expect him to stay in through that point as well, although honestly… who knows.
So… Super Tuesday. What does the picture look like then? Let’s assume that the RCP polls hold for South Carolina and Nevada (if they’re off, it’ll still probably be close). For the Super Tuesday votes, I’ve used RCP averages where available (Georgia, Massachusetts, Oklahoma, and Texas) and fallen back to the RCP national average for all of the rest (Alabama, Alaska, Arkansas, Minnesota, Vermont, Virginia, and Wyoming). Note that some of these states have poll numbers that are massively out of date, but they generally reflect the ordering of the national polls so I’ve stuck with them. This won’t be perfect methodology, but it’ll get us generally close. Using the RCP delegate simulator (and adjusting for the fact that its totals don’t quite match already allocated delegates), I get the following:
- Trump – 308 delegates and 13 state wins
- Cruz – 205 delegates and 2 state wins
- Rubio – 89 delegates
- Carson – 37 delegates
- Bush – 26 delegates
- Kasich – 24 delegates
That’s 689 delegates allocated – almost 28% of the total 2472 delegates. It’s also 15 out of the 56 total contests already “won.” Note that out of those, only two candidates (Trump & Cruz) have any wins counted at all and Trump already has enough to be entered for nomination.
Let’s make a few assumptions now. Let’s assume that Carson gets a clue and drops out at this point. Let’s also assume that Kasich drops due to lack of funding. Bush is a wild card. He may hold out until Florida votes on the 15th, or maybe he’ll drop, too.
It’s always hard to predict whom voters will see as their second choice. But here are my calls. Carson’s voters will predominantly continue to be anti-establishment. That means they’ll mostly split between Trump and Cruz, with a small but significant minority moving to establishment players. Let’s assume for the moment that Trump and Cruz split their share fairly evenly and that the minority predominantly goes for Bush. Let’s say 6 points for Carson becomes 2.5 Trump/2.5 Cruz/0.5 Rubio/0.5 Bush.
On the flip side, Kasich’s supporters will mostly go for the establishment “anyone but Cruz or Trump” lane. Let’s say that his 8.8% becomes 4% Bush, 4% Rubio, 0.4% Cruz and 0.4% Trump.
Assuming Bush stays in, we have the following situation in the national polls:
- Trump – 33.8% + 2.5% from Carson + 0.4% from Kasich = 36.7%
- Cruz – 21% + 2.5% from Carson + 0.4% from Kasich = 23.9%
- Rubio – 16.3% + 0.5% from Carson + 4% from Kasich = 20.8%
- Bush – 4.5% + 0.5% from Carson + 4% from Kasich = 9%
In this scenario, Trump’s hand has only gotten stronger. We could assume that Carson’s voters go much heavier for Cruz. But even if they do, it’s not enough to close that 13 point gap. Likewise, we can assume that Bush drops out as well and all of his support goes to Rubio. That’s enough to put him ahead of Cruz, but still not enough to close the 16 point gap between him and Trump.
Whatever Bush does, neither Cruz nor Rubio has the incentive to drop out yet. Both will have money and organization. And yet in the three way race, Trump continues to dominate. His plurality in the votes continues to be enough to win him a majority of delegates.
The only hope that Cruz, Rubio and Bush have is to survive the attrition to face a one-on-one match with Trump. The problem that each of them faces, however, is that none of their rivals (save perhaps Bush) has a particularly strong incentive to drop out and let them have that one-on-one race. The game here is one of chicken, primarily between Cruz and Rubio. Whichever of them holds on the longest has the best chance of beating Trump.
And yet the reality is that both of them are strongly incentivized to hold on long enough that it’s mathematically too late for the other to actually win the victory against Trump. By March 1, 27% of delegates will already have been awarded. By March 15th, when I project the other candidates will seriously think about dropping out, 62.5% of the delegates will have already been awarded. Even assuming that Cruz or Rubio survives to a mano y mano with Trump at this point, and even assuming that they carry the rest of the states with actual 51% plus majorities, it’s mathematically very difficult for them to actually win at this point. If Trump does well in New York (home turf plus a large haul of 95 delegates) and a handful of other states, there simply isn’t a path for Cruz or Rubio to a majority.
And those assumptions are both very strong. When do donors stop giving money to campaigns that aren’t winning? At a certain point, people stop throwing good money after bad. And for either candidate Mr. Silver’s contention about Trump also holds true. In a one on one race, Cruz or Rubio would also need to hit 51%. I think Mr. Silver would agree with me that Cruz’s chances of besting Trump one-on-one are middling at best. He polls well with evangelicals, but not any group outside of that. Rubio might do better. But even Mr. Silver – who has been thinking Rubio had the best shot for months – would have to agree at this point that it’s not a given that Rubio would hit 51% against Trump.
A key point to remember: by this point in the game, Trump’s ~35% plurality will have already carried him a far portion of the way to a majority of delegates. As I noted above, 62.5% of the delegates will have already been awarded, roughly half of which are likely to be Trump delegates. In order to win outright, his one-on-one challenger would have to essentially run the table. But even to deny Trump an outright majority and force a brokered convention, his challenger would have to do extremely well. They’d have to do much better than seems likely at this point. They’d also still have to win at least 8 states outright to even have their own name thrown in the hat.
Pulling off a mano-y-mano victory at such a late point, for either Mr. Cruz or Mr. Rubio, would be akin to an Superbowl team hitting the middle of the third quarter with a 21-0 deficit and managing to come back for a win. It’s theoretically possible. It’s no doubt happened in regular season games, and possibly even at a Superbowl. But nobody actually thinks it’s going to happen until it actually does. In presidential primaries it would be unprecedented.
Also consider that a brokered convention isn’t necessarily in the best interests of either Rubio or Cruz personally. There’s no guarantee at all that such a convention doesn’t just nominate Trump anyway. That’s not a given; if Trump has 40% or more of the delegates he needs and the convention doesn’t nominate him, expect a major voter revolt in the GOP. But even if a brokered convention doesn’t nominate Trump, there’s no guarantee that it would nominate Cruz or Rubio. Cruz isn’t well liked by the establishment and is unlikely to fare well at a brokered convention. Rubio is well liked, but anything could still happen – including but not limited to his mentor Jeb Bush having his name thrown in the hat.
Put yourself in their shoes. Would you continue a Quixotic campaign that was sure not to get you the delegates you needed to actually win just for the off chance of coming out ahead in a brokered convention? Or would you throw your support behind the front runner well before the convention in exchange for concessions – possibly including a VP slot? Trump literally wrote the book on cutting deals. Expect him to be ready to cut one here, if it becomes necessary.
The problem is more than simple delegate math. It’s game theory. And game theory predicts that neither Cruz nor Rubio will exit the race until it’s too late for the other to win. It further predicts that one or both of them is likely to eventually cut a deal with Trump to throw their support to him and finalize his delegate majority. Barring a sudden change in the polls, the nomination is Trump’s to lose. Like it or not, this is the truth on the ground. Mr. Rubio and Mr. Cruz know this, and by the end of March their campaign advisers will be reminding them of it heavily.
- Cruz Eliminated
- Trump Can Win With 30%
- South Carolina 2016
- Super Tuesday Post Mortem
- The Stupidest Thing Nate Silver Has Ever Said