The voters decided. Imagine that.

Photographer: Drew Angerer/Getty Images

My Mea Culpa on Trump Is Different

Jonathan Bernstein is a Bloomberg View columnist. He taught political science at the University of Texas at San Antonio and DePauw University and wrote A Plain Blog About Politics.
Read More.
a | A

Political scientists are going to be busy for years figuring out what happened in the 2016 Republican presidential race. But I'll try to deliver a nutshell version of what I think happened and how I, like many others, got it so wrong. 

The “party chooses” idea in political science basically makes two claims about the nomination process: That party actors will converge on a single candidate, and that they will exert sufficient influence over voters in primaries and caucuses to push this candidate to victory. In my view, the first part worked out normally this year, but the second part was a total flop.

Marco Rubio was the party’s candidate for the 2016 cycle. This was increasingly obvious from late fall 2015 on, and he wound up -- before his campaign fell apart -- with a clear lead in endorsements at various levels.

To be sure, he wasn’t the kind of party choice that Hillary Clinton was for the Democrats in this cycle. Instead, Rubio followed the pattern of John McCain in 2008 and John Kerry in 2004 -- lukewarm selections from a split field, with party actors waiting for evidence from the first caucuses and primaries to make their final decisions. 

In fact, by sticking with Rubio after Ted Cruz won in Iowa and Donald Trump took New Hampshire, the party actors demonstrated that they were using the early voting results as one input, not simply following whatever the voters wanted.  

Not only that, the Florida senator was the kind of coalition-style candidate who the “party chooses” researchers expect parties to select over factional candidates such as Trump, Cruz or Jeb Bush. (Factional candidates appeal to one or two party groups, but show little interest in the rest of the party.)

Why didn't the voters go along? 

This gets us to where the critics of the "party decides" theory (like Vox’s Andrew Prokop) were correct, and where I was dead wrong. Rubio, after doing well in Iowa, nosedived in New Hampshire after a poor debate performance. When he bounced back after middling second-place finishes in South Carolina and Nevada, it appeared as if voters would eventually follow party leaders’ cues as the field winnowed. But then Rubio fell short on Super Tuesday on March 1, and shortly thereafter his support cratered, despite continued strong backing from the party.

Many other political scientists and I thought parties could affect, or even dominate, the information environment, so that voters would mainly hear good things about the party’s chosen candidate and bad things about the contenders it strongly opposed.

I believed well into February that the saturation coverage for Donald Trump would give way to normal reporting once the primaries and caucuses began in earnest. That didn't happen. As a result, voters never heard the cues from the party -- information that was drowned out by the continuing focus on Trump. Even when some of that coverage was negative, it was still about Trump, meaning the other candidates never found their potential constituencies.

Another factor was the Republican dysfunction. I’d long said the Republican Party was broken, but unlike Norman Ornstein, I didn’t think it would affect the nomination. After all, it hadn’t in 2012 and 2008. This time was different, as Ornstein predicted.

Some of the most visible Republican party actors -- talk-show hosts and others in the Republican-aligned media -- were for Trump or at least willing to give him plenty of positive publicity, despite his many violations of conservative orthodoxy. This diluted the message that Rubio was the party’s pick. Of course, there is often dissent among party actors even after the bulk of them have settled on a candidate. What was different this time is that the dissent from some of the loudest voices in the party, such as Sean Hannity, obscured the consensus choice. 

Related to the dysfunction is that for years Republican figures were bashing the party's “establishment” and supposed Republicans In Name Only, a barrage that may have made rank-and-file voters unusually resistant to cues from their party. Of course, no one ever believed that armies of voters simply did whatever their leaders told them to do. But this time a significant number of voters may have actively wanted to oppose candidates if they had “establishment” backing.

So put all of that together, and information just didn’t work in the way I expected. I should have seen some signs of that from the beginning and certainly as time went by. 

But even if this was a necessary component to Trump's victory, it was probably not sufficient. There are a whole string of other contingencies as well. If Rubio had come from a state that voted earlier in the process. If Rubio hadn’t frozen in the New Hampshire debate. If John Kasich had left the race when he was beaten. If Rubio’s attack-dog act had been more successful. If someone other than Cruz -- hated by many in the party and probably not appealing to voters either -- had been the final opponent for Trump. If Republican delegate allocation was just a bit more proportional, or if Republicans had superdelegates as the Democrats do.

Trump didn’t win by much. Therefore, he needed all or almost all of those “ifs” to fall his way.  

The main problem, however, was in translating the party's signals into actual primary and caucus votes. What’s less clear, and will take more research, is which factors played the greatest role in the Republican breakdown -- and whether the “party chooses” model will still be a useful framework in future presidential contests. 

  1. The most well-known work is Cohen et al.’s "The Party Decides." I’ve contributed to the literature as well.

  2. In fact, Rubio dominated public endorsements from fall 2015 on, but the pace really picked up after Iowa and New Hampshire. 

  3. Similar to CNN's "missing airplane" coverage, and supplemented, especially online, by coverage driven by analysis of what audiences were watching and clicking on. Of course, news coverage has always been affected by what was popular, but until recently no one really knew in detail what viewers and readers were interested in.

  4. Remember, it’s always true in presidential campaigns that most candidates never find their potential constituencies. So saying that Trump was just inherently appealing to Republican voters isn’t an explanation of how he won. Each of the 20-some candidates who made at least initial efforts had some appeal to many Republicans. 

  5. Nate Silver deserves more credit for consistently arguing that many of us underestimate the chances of unusual outcomes. 

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

To contact the author of this story:
Jonathan Bernstein at

To contact the editor responsible for this story:
Katy Roberts at