By David Byler – – – –
The smoke-filled room is back.
Or at least that’s what it looked like last week when Senate Majority Leader Mitch McConnell, RNC Chairman Reince Priebus and about two dozen other high-ranking Republicans met in Washington to discuss the possibility of a contested convention. Some of these leaders are worried that when the convention rolls around next July, no GOP candidate will have amassed enough delegates to win the nomination – giving current frontrunner Donald Trump the opportunity to wheel and deal his way to the nod.
So how worried should these establishment Republicans be? It’s tough to put a hard probability on a brokered convention occurring, especially when such a probability will change as we move through the primary calendar. But one key factor is candidates dropping out. Specifically, which candidates do so — and when — could have an outsized impact on the result of the primary.
When Will Candidates Start to Drop Out?
It’s impossible to know for sure, but looking at past presidential races gives us some clues.
This graphic – similar to one created by the New York Times’ Alicia Parlapiano – shows when candidates entered and left the presidential race in crowded primaries since 1992 – namely the Republican contests in 2016, 2012, 2008, 2000 and 1996 and the Democratic races in 2008, 2004 and 1992. The blue lines represent Democrats and the red lines represent Republicans. (Data for four minor campaigns – Herman Cain’s 2000 Republican bid, Robert Dornan’s and Alan Keyes’ 1996 Republican bids, and Eugene McCarthy’s 1992 Democratic run – were unavailable.)
The curve for the 2016 GOP primary towers higher than the rest and stops 30 days before Iowa (other candidates might drop out before then, but we’ll assume they won’t for the purpose of this visualization). This cycle’s race is extremely crowded, but so far the overall pattern has been similar to that of past primaries – candidates got in sometime in the year leading up to the Iowa caucuses, a few have dropped out, but most have stayed in.
In past elections, candidates have quickly dropped out after faring poorly in Iowa, New Hampshire and other early contests. We’ll likely see a large number of dropouts after those primaries again in 2016, but note that the exact rate of dropouts differs from year to year. In some cases – as in the 1996 Republican cycle and the 2004 Democratic cycle – a consensus candidate emerged, and most of his chief competitors left the race, within about two months of Iowa. But in other contests – such as the 2012 Republican race – the primary stretched longer, and it took more time for the field to narrow to two competitors or the eventual victor.
Why Lingering Candidates Might Make Such a Big Difference
This is relevant because the calendar has clear ideological divisions, and whether the field stays crowded weeks into the primary season or months into it makes a big difference. Other analysts have made similar observations, but I think this graphic drives the point home.
Here are the details to explain what this rather complicated graphic shows: Each point is a state or congressional district, and the size of the point is proportional to how many delegates are available in that contest. The horizontal axis is the date of the voting, and the vertical position (and color) of each point shows the partisan lean of the state or district in the 2012 presidential election. For states, the lean is calculated by subtracting the Republican two-party share of the national presidential vote from the Republican two-party share of the state’s popular vote (all in decimal form). So a value of 0.1 means that the state was 10 percent more Republican than the national average and a value of -0.1 means that it was 10 percent more Democratic than the national popular vote.
The same formula was used for congressional districts, with the presidential vote in each district helpfully calculated by Daily Kos Elections. Partisan lean isn’t a perfect measure – it makes Iowa, which has a very conservative Republican primary electorate, seem centrist – but in most cases it’s a useful estimate. The points with a black outline are winner-take-all states or districts and the rest are deemed proportional (that is, there’s a way for candidates to split the delegates). Also note that the data here are from 48 states plus Washington, D.C. (Colorado and Wyoming don’t have official primary votes, so they’re left out. Territories are omitted because they don’t vote in presidential general elections.) Note that states often allocate delegates to the winners of each congressional district – thus the small dots throughout the graphic.
To see how much dropout order matters, consider a few scenarios (RCP Senior Elections Analyst Sean Trende and I briefly discussed some of these scenarios here).
Maybe the best scenario for the GOP establishment would be one where a large number of conservative candidates stay in through March, while only one establishment candidate remains. For instance, imagine a scenario where Donald Trump, Ben Carson, Ted Cruz and Chris Christie make it to the March primaries. Many Southern, highly evangelical red states vote in early March, so the three conservative candidates (I’m using this term broadly, since Trump’s support is more moderate on some issues, and tends to be secular and anti-establishment) could split that large red mass of delegates near the beginning of the calendar. Christie might win a couple more liberal states like Minnesota and Massachusetts while quietly cleaning up in some of the more liberal districts within those red states. When Ohio and Florida (large, moderate-leaning winner-take-all states) vote in mid-March, that will provide another opportunity for a more establishment-leaning candidate to grab delegates. And from there on out the conservatives have very little friendly territory to mine, while the establishment candidate has many blue states and districts stretching across the rest of the calendar. In that case, the early fracturing of the conservative vote and consolidation of the establishment would create a big advantage for the more centrist candidate and thus decrease the chances of a contested convention.
Now imagine the reverse scenario – there’s one ascendant conservative (say, Cruz) and multiple establishment candidates (Rubio, Jeb Bush and John Kasich) survive the carve-out states. This scenario would give Cruz a big advantage. If he were able to grab most of the delegates from that large mass of conservative states in early March while the establishment candidates split the left-leaning states and districts, Cruz could go into Ohio and Florida with a large lead. If no establishment candidate clearly won out in mid-March (e.g. Rubio wins Florida but loses Ohio to Kasich), and the conservative candidate gets more delegates from Utah and Arizona (the big red states voting prior to April), this could create a scenario where the establishment resigns itself to an ascendant conservative opponent. Alternatively, it’s possible that the establishment coalesces around one person in late March to mid-April when there are relatively few contests, and then that candidate quickly racks up a large number of delegates from the remaining blue-leaning states. In that case, a contested convention or a late win by the conservative candidate are both plausible outcomes.
Another possibility is that some winnowing happens, but a relatively large number of establishment and conservative candidates stay in through March. In an ordinary cycle this would be unlikely – donors would withhold money from losing candidates and dropouts would come more quickly. And this cycle might turn out to be ordinary – but we aren’t yet sure of that. Maybe the functionally limitless money from super PACs helps candidates survive longer than they would have in previous cycles. Maybe the size of the field matters. After all, if the carve-out states knock out eight candidates, six would still remain. In either case, if multiple candidates divvied up the March delegates and stubbornly refused to drop out, then the likelihood of a contested convention would seem to increase.
Finally, it’s possible that one candidate emerges from each wing of the party after the early states, and the two end up in a Clinton-Obama-esque showdown. Suppose that those candidates are Rubio and Cruz. In that case, Cruz could take an early delegate lead from the conservative states and Rubio could rack up delegates from more liberal states later in the calendar. This would leave open the possibility of an exciting photo finish, but the likelihood of a contested convention would drop, as getting to 50 percent plus one is easier when multiple candidates aren’t divvying up the electorate.