Pundits have assured us that the support for Donald Trump is so limited that he can’t possibly get the GOP presidential nomination. Last week in The New York Times, Ross Douthat argued that Trump has a ceiling around 30 percent of Republican voters and consequently will be defeated. To put this numerical claim to the test, I have created a detailed state-by-state simulation of the nomination rules. My conclusion may surprise you: Trump’s current level of support may be enough to deliver him the nomination on the first ballot at the Republican National Convention in July.
Political science research suggests that the “party decides,”—that is, a party’s insiders ultimately have a decisive say in choosing the presidential ticket. In this year’s splintered Republican field, however, the power players have been largely silent, as if waiting for a “mainstream” leader to emerge. The frontrunner in polls, Trump, has never held office. In second place is Senator Ted Cruz, an outsider to his own party’s inner circles. Other conventional candidates (current and past officeholders) have combined support totaling about 25 percent. Have the insiders waited too long to consolidate behind one candidate?
The Republican Party’s nomination process has a simple, basic premise: If any candidate arrives at the July convention in Cleveland with a majority of the delegates, he (it’s going to be a he this year) can win the nomination on the first ballot. Delegates are assigned, however, through a byzantine process—one that does not require the eventual nominee to get support from a majority of voters.
This feature of the process provides Trump with an opportunity. In surveys of the first three states to vote—Iowa, New Hampshire, and South Carolina—Trump’s support is now between 27 percent and 31 percent. How do these numbers translate into delegates?
For historical guidance, we can look to the 2012 Republican nomination race. Early phases of that four-way race also had a frontrunner who lacked majority backing: Mitt Romney. In the first 22 primaries and caucuses ending on Super Tuesday, Romney received an average of 40 percent of the popular vote—yet he won 59 percent of the delegates.
To see how the 2012 primaries dealt with a divided field, I looked up the outcome of each binding caucus or primary leading up to Super Tuesday. This graph shows the vote share and delegate share of first-place finishers. The black diagonal line indicates what would happen if delegates were awarded in proportion to vote share.
The graph shows only those delegations that were determined by rules similar to this year’s, and excludes those chosen under rules no longer in effect. For example, Ron Paul supporters in 2012 found ways to seize control of delegations from other candidates after the election in battles that sometimes went to the Republican Party’s credentials committee. For 2016, the Republican National Committee has closed off this possibility by making primaries binding upon delegates.
In nearly every case in 2012, vote shares of 35 percent or higher were associated with a greater-than-proportional share of delegates. This pattern may come as a surprise, since on the surface, Republican party rules specify that early primaries and caucuses should not be winner-take-all, but proportional. But my close study of the rules reveals a rather different story.
Understanding the rules was quite a job; imagine preparing tax returns in all 50 states at once, and you get the idea. Here is part of one specimen, which governs how 47 of Texas’s 155 delegates are allocated:
- If a candidate receives a majority of the vote (more than 50 percent) that candidate is allocated all 47 at-large delegates. [Rule 38. Section 9.a. and 9.b.]
- If no candidate receives a majority of the vote and at least 2 candidates receive 20 percent or more of the vote, the 47 at-large delegates are allocated proportionally among those candidates receiving 20 percent or more of the vote. Rounding rules: Beginning with the candidate receiving the largest number of votes, round any fraction to the next whole number of delegates. Continue this process with the next highest vote getter and repeat until all the delegates are allocated. [Rule 38. Section 9.b.]
Eventually I noticed two features of delegate selection that work against proportional outcomes:
1) At a statewide level, a minimum level of support is required to get any delegates at all. Up through Super Tuesday (March 1 this year) the threshold for representation is as high as 20 percent, with a median value of 15 percent. In current polling, only Trump consistently rises above this cruel cutoff. Ted Cruz is near the threshold, while Senator Marco Rubio and others are likely to be cut off at the knees.
2) Other delegates are assigned locally, three per congressional district. In addition to thresholded representation, districts also often award delegates only to the top two finishers. Currently, this rule favors Trump and Cruz. The rest get bupkis.
Expressed as graphs, here are some typical-looking state and district-level rules:
The graph on the left shows a statewide rule that applies in Alabama, Georgia, Texas, and Vermont. In these states, a candidate who gets a majority vote wins all the statewide delegates. Otherwise, delegates are split up proportionally among candidates who make it above the threshold. Because candidates below the threshold are left out, delegate shares are larger than vote shares. The graph on the right shows the district-level rule in Alabama, Georgia, and Texas. The first-place finisher gets all three delegates if he wins over 50 percent of a district’s vote; otherwise, he gets two delegates.
Sometimes inequity is compounded in other ways. In New Hampshire, for example, those who finish above a 10 percent threshold get that numerical share of the state’s 20 delegates—but leftover delegates go to the top finisher. Under current conditions, Cruz, Governor Chris Christie, and Governor John Kasich would get two delegates each, Rubio would get three—and Trump would get 11.
These rules are representative of a broader process that is documented in detail at The Green Papers. I have written a computer program to simulate delegate selection from now until Super Tuesday, a process that commits 727 of the nomination process’s 2,472 delegates. I used this simulation to evaluate a 2012-like scenario in which the field is winnowed down after Iowa and New Hampshire to four candidates, with the support of dropouts going to the survivors. I also plugged in 2012 levels of state-to-state variability, which were similar for all four top candidates (each one had a standard deviation of 12 to 13 percentage points). The result of these calculations is an output that plausibly reflects how any frontrunner would do when faced with three other candidates. For example, it might show us how Trump would fare when trailed by Cruz, Rubio, and former Governor Jeb Bush as his remaining opponents.
Here is what the frontrunner’s average delegate share looks like, when graphed against his levels of pre-winnowing support—that is, in terms of polling data that is available today.
What this analysis shows is that in a divided field of candidates, a candidate polling at 30 percent or above before Iowa and New Hampshire might reasonably expect to win 50 percent of the delegates awarded through Super Tuesday, an initial step toward an overall majority. Today, Donald Trump meets this criterion.
To maintain his route to an outright majority of delegates after Iowa and New Hampshire, Trump’s support in primaries and caucuses will need to rise to around 40 percent as lower-tier candidates drop out. This number serves as a second useful indicator to predict his eventual fate at the Republican National Convention. In terms of support, it would mean that Donald Trump is this year's Mitt Romney.
The mirror image to Trump's advantage is Senator Marco Rubio's disadvantage. In my simulations, Rubio’s current support translates to about 10 percent of delegates, even after he picks up votes from a narrowing field. This outcome is expected whenever four candidates survive. Fortunately for Trump, the primary calendar is likely to encourage several of his opponents to stay in the race at least until March 15. On that date, Ohio and Florida will have winner-take-all primaries, creating an incentive for favorite sons Kasich (who is from Ohio) as well as Bush and Rubio (who are from Florida) to stay in even if they are losing. Winning their home state would give them power at the convention if no candidate has a majority of delegates. Florida is a good example of the problem: In polls, Rubio's current support (16 percent) and Bush's (11 percent) put them far behind Trump at 33 percent. To recall the old Ben Franklin saying, since Bush and Rubio won’t hang together, they may hang separately.
Although my calculations show the advantages that Trump currently enjoys, his nomination is by no means assured. If the field does narrow to Trump, Cruz, and Rubio, Rubio would likely accumulate enough support to be a close second to Trump. To make a large impact, this narrowing would have to take place well before Super Tuesday. The window of opportunity for the Republican “establishment” is closing fast.
Finally, if no candidate gets to an outright majority, the convention becomes genuinely suspenseful. Party insiders should not necessarily be consoled by this idea. Delegates are usually selected for loyalty to their candidate. If current trends were to persist, the convention floor in Cleveland would be filled with close to 1,000 Trump delegates. These delegates won’t be from the usual pool of party loyalists. They seem like an unpromising starting point for elites to work their magic.