What I Got Wrong About the 2016 Election

Political scientists like me favor the "party decides" theory of presidential primary selection. It didn't work for Republicans.

Who knew?

Photographer: JOHN GURZINSKI/AFP/Getty Images

Time for the yearly exercise in self-shaming. 

The most obvious one this year is one on which I was very, very wrong indeed: My insistence that Donald Trump would not be the Republican nominee for president. In my defense, almost everyone got this wrong at one point or another ... but that's no excuse, and I was more confident in my wrongness and stuck with it longer than most. 1

QuickTake How the U.S. Elects Presidents

I got it wrong because I relied on the conclusion some political scientists, myself included, had reached: That political parties, including their extended networks, were responsible for the selection of each presidential nominee beginning in the 1980s, and that they did so by reaching agreement -- which voters in caucuses and primaries then ratified.

The first part of that process -- party actors reaching a decision -- worked this time for Republicans when many of them converged, albeit late and weakly, on Senator Marco Rubio of Florida. The second part didn't work at all; Rubio fizzled out early. 

For the first part, I think the "party decides" theory was always underdeveloped -- we needed a better explanation for close contests such as the Democrats' nominating battles in 1988 and 2004 and the Republican ones in 2008 and 2012. I think that goes beyond current problems within the Republican Party. 2  

But it was the second part that went most badly astray this time, in my (very much after-the-fact) view. What happened? We assumed media norms -- that is, patterns of behavior -- which had held in the past would always hold. Television news, for example, never gives equal time to all candidates, but it usually moves from hyping one candidate to hyping another one, so multiple candidates wind up with exposure. Most media outlets usually tend to give extra attention to party-favored candidates over those without endorsements (as Hillary Clinton received far more coverage in 2015 than her opponents). Conversely, when the primaries and caucuses begin, the media tends to downplay the chances of the leading candidate and play up the chances of one or more trailing candidate.  

Instead, television news, led by CNN, chose to give saturation coverage during the nomination phase to a single candidate, one the Republican party had no use for at all. 3  To be sure, much of that coverage was critical, but what turned out to really matter was that the other candidates were unable to break through enough to stand a chance. 

For my part, I was probably wrong to treat "party decides" conclusions with as much confidence as I did, and I was definitely wrong to be slow to realize media norms had evolved and why that was important. 

I think I was a better guide to the general election in most respects. But I was at least partly wrong about one thing: My claim that partisans would wind up loving their party's nominee, even if they supported another candidate in the primaries. 

To a large extent, that did happen as it usually does. A lot of Democrats who supported Hillary Clinton reluctantly in the spring wound up enthusiastic about her in the fall, and a lot of Republicans who wanted nothing to do with Trump in November 2015 found redeeming qualities in him by November 2016. That a fairly large minority of Republicans never warmed to Trump was no surprise, given how unusual a candidate he was. But it remained true that a surprising (albeit still fairly small) number of Democrats remained lukewarm or worse towards Clinton in the weeks leading up to the election, a sentiment later symbolized by a record number of electoral college defections

Remember, this isn't about why Republicans or even independents didn't like Clinton. It's about a small, but apparently bigger-than-I-expected group of liberal Democrats resisting the normal campaign pressure to fall in love with their party's candidate. And I'm not sure why. Was it a media story, something about the continued skepticism among elite reporters about Hillary (and Bill) Clinton? Some remnant of gender bias among some Democratic voters? The still-remembered vote she cast for the Iraq War? A sign that at least one small faction of Democrats may be becoming as dysfunctional as Republicans in refusing to accept imperfect (that is, all) candidates? I don't know. I'm pretty sure it wasn't about the specific details about Clinton's poor information-management choices -- partisans have embraced nominees with far more obvious liabilities. 

I think these were the two biggest things I got wrong in 2016. I'll try to do better in 2017, and in the (gulp) just beginning 2020 presidential nomination contest. 

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
  1. A better excuse is that Trump's win was in fact a very close one, and probably depended on a whole lot of unlikely things happening. I do believe that's what happened, but it's still not much of an excuse -- especially because I was (at times) so confident in my incorrect analysis. 

  2. A quick note for specialists: I don't think it's sufficient to revert back to Nelson Polsby's conclusion that in the absence of parties nominations devolve into random affairs heavily influenced by unpredictable media effects. The Republican Party was not absent in 2016, the way the Democratic Party was (mostly) absent in 1976. 

  3. To be clear: I'm not "blaming" CNN and the others here. I'm saying that the way they acted in 2016 on the Republican side was different from the way they acted in previous cycles, for better or worse -- and that "party decides" conclusions had been based on an assumption, now proven wrong, that the old media behaviors would continue. 

To contact the author of this story:
Jonathan Bernstein at

To contact the editor responsible for this story:
Jonathan Landman at

Before it's here, it's on the Bloomberg Terminal.