It is a major problem for our public discourse that pundits are rarely held accountable for their mistakes, and they hold themselves accountable even less. I decided to audit my own election coverage–I went “on record” seven times to make predictions about the race. Here were the results:
I. August 2015, Salon.
Predictions: Progressives seemed to be growing increasingly hostile and condescending towards working-class white voters—minimizing their grievances, villainizing them as racist or sexist, condescending to them about how they supposedly vote against their interests. I argued that not only is this approach non-productive for realizing the policies activists ostensibly aspire towards—but if they kept it up with this militant parochialism, working-class whites could rise up and kill the entire social justice movement.
Outcome: Progressives not only perpetuated, but upped the ante on their narratives—portraying virtually all opposition to Hillary Clinton as arising from ignorance, sexism or xenophobia. The result: Democrats got the lowest share of the white vote since 2008, to include mid-term races.
II. October 2015, Salon.
Predictions: The Republican Party seemed to be gearing up for a “law and order” campaign a la Nixon. But Democrats would be foolish to rely on a Republican boogeyman to ensure black turnout and allegiance.
Outcome: Upon winning the nomination, Donald Trump did run a law and order campaign. And the Democrats did rely heavily on Trump’s negatives to drive minorities to the polls, rather than emphasizing a concrete and positive set of proposals which would improve black lives and plausibly be passed in a Clinton Administration. The result? Blacks turned out less than they did for Obama–and of those who showed up, the largest share since 2004 voted for the Republican candidate instead.
III. March 2016, Huffington Post
Predictions: If the Democrats nominated Hillary Clinton, she would lose to Donald Trump. The polling at the time which a Clinton lead was premature—in part because many voters were still holding out hope for Sanders, Cruz or Kasich. But when it just came down to the stark choice of Trump and Clinton, the public would break for Trump. Disavowals by prominent Republicans would be irrelevant to the base, who largely despise these establishment figures.
What would Trump’s victory map look like? From the data on hand at the time, of the states held to be “in play,” I suggested Trump would win Florida, North Carolina, Georgia, Arizona, Ohio, New Hampshire—and possibly Wisconsin, Michigan and Nevada (I didn’t make a prediction about Pennsylvania either way, because I didn’t have enough information. But I had enough to see the outcome of the general election).
Data: This was a central prediction (and also my first essay on domestic politics outside of racial issues or national security contexts). For this reason, I’ll explain my method a little more. The first piece of the argument looked at highly-competitive states where Clinton lost but Trump won. These states, I argued, suggested real vulnerability for Clinton in the general election. Swing states where both Clinton and Trump performed well could be more competitive in principle. Looking at the map in this way suggested Clinton may be in trouble–in any case, the confidence she and her followers seemed to have was not well-founded.
I highlighted other signals that seemed to reinforce this worry: relatively low turnout and enthusiasm among Democrats in these states, and high-turnout for Republicans. New registrations and party-realignments in these states seemed to favor the Republicans. Polling, while not yet predictive vis a vis the final outcome of the general election (for reasons I detailed in the article), did suggest that Clinton was more vulnerable to Republicans than Sen. Sanders–and they hadn’t even really gotten warmed up in attacking her yet (but predictably, scandals loomed).
When these considerations were combined, it seemed evident that nominating Hillary Clinton was a mistake. Especially by means of a process which would alienate and demoralize essential Democratic voters (such as by having the super-delegates contradict the clear will of the people in their states, or having them make known their pro-Clinton preferences prior to most voting).
Outcome: Hillary Clinton was nominated by the DNC. Trump won all of the states I mentioned with the exception of Nevada and New Hampshire–to include left-leaning Wisconsin and Michigan, a possibility team Clinton had laughed off as delusional. Democratic plays for Georgia, Arizona, Utah and other right-leaning states were in vain, as anticipated.
As predicted, Republican elites spurned Trump, and it hurt them more than him. And as people increasingly turned away from the third-party options, they broke for Trump instead of Clinton. It was a highly-specific and highly- prescient set of predictions, which was met with a lot of derision at the time from the Clinton camp.
IV. May 2016, Salon
Predictions: Contrary to the prevailing narratives at the time that the race was a lock for Clinton, and that Trump was a uniquely weak opponent for her—I took the position that the advantage seemed to lie with Trump, and that he was actually better positioned than any other candidate to repudiate Hillary Clinton and what she represents.
I highlighted that Democrats rely on irregular voters to win national contests, but these voters are unlikely to turn out if the race becomes a “lesser of two evils” match. I proposed Clinton’s numbers would only go down, while Trump could gain on her in many critical respects (such as trust). I argued that painting Trump as a bully, misogynist or racist would not peel white voters away from Trump, but instead rally them around him more firmly—to include white female voters. I argued that Trump appealed to far more than poor, uneducated white men. I predicted that Clinton’s advantage in fundraising, infrastructure and analytics was not going to be decisive in this campaign.
Outcome: Clinton relied on negative partisanship and experienced low turnout of critical constituencies. Her attempts to paint Trump as a bully, misogynist or racist had the predicted galvanizing effect on the white vote–including turning many white-dominated regions that voted for Obama in 2012 towards Trump. Clinton developed a huge edge in fundraising, campaign structure and data mining; Trump had little organization to speak of, but he won anyway because he was great at defining the narrative and dominating the news cycle, and he fearlessly engaged with the public at every opportunity—as opposed to Clinton who virtually disappeared for more than a month while courting the super-rich.
V. June 2016, Huffington Post
Predictions: Despite narratives that Trump had a “white people problem,” I predicted he would win the majority of the white female and white college-educated vote. And moreover, despite narratives that Trump is a racist or xenophobe, he would win larger shares of black, Latino and Asian voters than Mitt Romney did in 2012. I argued that if he locked in these demographics, the Democrats will certainly lose.
Outcome: Trump did win the majority of the white vote—to include the lion’s share of college-educated whites, and white women. The Democrats made a huge drive to register and mobilize Latinos, but ironically, many of these may have ended up voting Republican: Trump got a larger share of the Latino and Hispanic vote than Mitt Romney did in 2012—a feat which most others viewed as utterly impossible at the time. Trump also managed to exceed Romney’s performance with blacks and Asian Americans, not to mention his gains among young people and registered Democrats and Independents over the 2012 numbers. The only real surprise was that, despite his groundbreaking homonationalist approach to LGBTQ Americans, he actually underperformed among this demographic relative to Mitt Romney.
VI. August 2016, Huffington Post
Predictions: Despite having championing Trump’s prospects for more than a year, I had a sudden lack of faith as crunch-time approached. In particular, I was disturbed by Trump’s dramatic and unnecessary rightward pivot, and the sudden turn in the media to being rabidly anti-Trump and pro-Hillary.
Outcome: Of course, Trump did win the election. In particular, the major media shift against Trump and in favor of Hillary helped reinforce discourses that the media was untrustworthy and partisan, that they are trying to hand Clinton the election, that there is an Establishment conspiracy to deny the will of the people, etc. Each new endorsement by major media outlets only reinforced this narrative: they didn’t undermine Trump, they undermined their own perceived credibility and objectivity. As I began to realize that this desperate media blitz against Trump was more likely to backfire rather than tank him, I shifted back to my initial and sustained conviction that Trump would take it all.
VII. September 2016, American Conservative
Predictions: In the conclusion of this article I argued that the antipathy beltway Republicans have shown to Trump will come back to haunt them—because he may well win not just despite, but because of his tensions with the establishment. And to the extent that he shuns the mainstream in turn, he’ll be left with a set of radical and ill-informed, but loyal and trusted, advisors to inform his policies. This will be bad for all of us. If Republicans are truly concerned about the damage Trump could do in office, washing their hands of him is a pretty dumb response.
Outcome: Trump did win, and precisely due to his perceived challenge to “business as usual” in Washington. He is set to appoint a strange set of characters to his cabinet, and it is likely that many of the Republicans who turned on him will be left out in the cold for the foreseeable future. Even more ironic: the reason for this betrayal by establishment Republicans was, in many instances, a concern about possible pernicious effects on down-ballot races (under the assumption he would likely lose the general election). Not only did the RNC maintain its hold on both chambers of Congress, in many instances Trump proved an asset to down-ballot candidates in highly-contested races. So they turned on Trump for naught, and they will likely pay a steep price for it. As may we all.
Conclusion: November 2016, The National Interest
In the aftermath of the election, I explain the major things I noticed which most analysts seemed to miss. The most important insight, recognized early, was that we were in an extraordinary cycle, and the normal rules of the game did not seem to be holding. Therefore, the usual predictors may not be as reliable as they typically are, and it may be necessary to look for other indicators which more reliably track the peculiarities of this race. Or perhaps to alternative models, such as professor Lichtman’s “13 Keys” approach or Norpoth’s primaries-based model (shout-out to my fellow dissenters!). I also discuss the limitations of polling, and how to partially compensate for these.