Nate Silver is one of the best things ever to happen to political journalism in America. He pushed journalists to rely on more than just vague hunches, to take poll results and other data seriously, and in the process raised the level of the profession immensely. My disagreement with his current analysis is not meant to take away from that contribution.
Yet in analyzing the chances for Donald Trump’s nomination, Silver has left the world of numbers behind and is instead focusing on emotion and on what parties normally do in similar situations. Sadly, as he became the establishment, going from writing his own blog to working for Disney, he began to think like them. This has led him to ignore what is right in front of him about Trump and to not acknowledge fully mistakes he has already made with respect to Trump. In August, Silver gave Trump a 2 percent chance of winning the Republican nomination. He has now upped the odds to 14 percent, but Silver’s expressed skepticism remains as strong as if this change was never made. It appears that being so wrong in August about Trump, Silver can’t quite bring himself to accept Trump’s strengths.
This article will refute many of the points made in Silver’s two leading Trump posts, "Donald Trump's Six Stages Of Doom" and "Three Theories Of Donald Trump's Rise."
Here’s a key point: A candidate who wins New Hampshire always has a real shot at the nomination. In the modern era, only three Republican candidates have won the New Hampshire primary and gone on to lose the nomination: John McCain in 2000, Pat Buchanan 1996, and Henry Cabot Lodge in 1964.
With New Hampshire less than three weeks away, Donald Trump seems to be in a fantastic position to win the New Hampshire primary. The RCP average has Donald Trump leading by an average of 17.7 points. If Trump lost half of his support in New Hampshire when the votes are actually counted, his level of support would still exceed that of any other candidate. In other words, he would still win.
And how likely is it that Trump will lose half his current level of support in the next two and half weeks? According to Public Policy Polling, Trump has the most heavily committed supporters of any of the candidates; 68 percent of those supporting Trump are considered firm in their support.
Trump is the clear favorite in New Hampshire, and a New Hampshire win gives him a strong shot at the nomination — at least historically.
So if Trump is doing so well in New Hampshire, how is it possible that Silver gives him only a 14 percent shot at the nomination? The answer is that Silver has to downgrade Trump’s chances in New Hampshire, which is what he does. Silver considers Trump “potentially vulnerable” in New Hampshire: “Trump could lose New Hampshire either to a surging Cruz or if one of the several establishment candidates — Marco Rubio, Chris Christie, Jeb Bush, John Kasich — can consolidate the support of more moderate/establishment Republican voters.”
Silver quantified his prediction using two models, one of which relies on polls and endorsements, and the other that only uses polls. (For those of you who want to dig into how Silver created his models, here it is.) When Silver relies only on polls, he gives Trump a 57 percent chance of winning New Hampshire. (Because Silver doesn’t explain exactly what goes into this model, it’s hard to evaluate it, though my sense is that Trump’s chances are still too low.) Adding in endorsements from major elected officeholders of which Trump has secured none, reduces Trump’s odds to 40 percent. Leaving aside the problems with the endorsement-plus model, which punishes Trump for not being popular with governors and Congress people (I’ve discussed this before on this blog), Silver still needs to explain how a 40 percent chance of winning in New Hampshire translates into a 14 percent chance of winning the nomination, given New Hampshire’s track record in picking the ultimate nominee. Well, it doesn’t if you look at the numbers.
Here’s an alternative theory of how the race is likely to shake out. Let’s assume that the standings in Iowa hold and Cruz wins while Trump comes in second (though the current polls give Trump a decent shot at winning there.) Trump wins New Hampshire fairly easily. This brings the battle between the two winners to South Carolina where the contest will likely be fierce. So far, however, Trump has a strong lead in that state, 14 percent in the most recent poll. Only one candidate has ever won South Carolina on the Republican side and gone on to lose. (Newt Gingrich in 2012.) No Republican candidate has ever won both South Carolina and New Hampshire and lost. The odds of Trump winning both states at this point seem to be around 50 percent.
Even if you accept Silver’s own models, and factor in a win’s effect on the coming contest (as Silver’s does), Trump should have about a 30 percent chance to win both New Hampshire and South Carolina based on the poll-plus model and an at least 40 percent chance on the polls-only model.
Even using his own polls-plus model, Silver would come up with a number much closer to Trump having a 25 percent chance of winning the nomination than his current 14 percent. Given that he started at 2 percent, it seems that Silver is refusing to let the numbers move his logic.
Part Two will explain how Trump can conquer all the stages of Doom.
The Scorecard
The Scorecard is a political strategy and analysis blog. Our hope is to provide information and insight that can be found nowhere else into how and why things are happening in American politics. Unlike many political pundits, we will tell you who we think is going to win as an election approaches; we will tell you why; and we will give you a sense of our level of confidence. Ours is a holistic approach, one that takes in as many numbers as possible but is also willing to look past the numbers if need be. When we turn out to have been wrong, we will let you know. When we are right, we’ll let you know that too.
Our Delegate Math
About Me
Author Jason Paul is a longtime political operative who got his start as an intern in 2002. He has been a political forecaster for almost as long. He won the 2006 Swing State Project election prediction contest and has won two other local contests. He had the pulse of Obama-Clinton race in 2008 and has been as good as anyone at delegate math in the 2016 race. He looks forwards to providing quality coverage for the remainder of the 2016 race.
0 comments:
Post a Comment