So, I suppose it is time for some political science reminders. First, we have no bloody clue how this works. Sure, I can tell you the rules. States have primaries and caucuses, which select delegates, who vote at the Convention, and a majority is needed, so if nobody gets a majority, you get a brokered convention, which won't happen, and blah, blah, blah. What I mean when I say that we don't know how this works is that we don't have a good predictive model, and never really did. Our general election models work. Yes, they worked in 2016. It was the polls that failed. Our forecasting models predicted a Republican victory. See Alan Abramowitz. We just don't have a good model for presidential nomination contests.
Why not?
Well, the big factor is lack of observations. Think about it. The rules of the game changed after 1968. In 1968, the Democratic Convention broke down in riots because Eugene McCarthy won the primaries, but those primaries didn't matter. The muckety-mucks at the Convention chose to nominate Hubert Humphrey anyway. Riots ensued, and Humphrey lost to Nixon. In the aftermath, the Democrats formed the McGovern-Fraser Commission to write a new set of rules. Those rules are the rules under which delegates are selected by primaries or caucuses rather than just being good, little party muckety-mucks. So, we only have relevant data going back through 1972.
OK, but general elections are only comparable going back a bit further, so what's the big deal?
Well, incumbents are always re-nominated. (There's an interesting question about what would have happened in '68 if Johnson had kept going, but that's pre-McGovern-Fraser, so it's also irrelevant from our perspective.) Yeah, you'll have some interesting contests, and they can be nasty, but every incumbent in the MF-er era, as I'm going to call it in this post, has been re-nominated. So, Nixon in '72, Ford in '76, Carter in '80, Reagan in '84, Bush the Elder in '92, Clinton in '96, Bush the Younger in '04, Obama in '12... uninformative contests. Sitting VPs? Complicated. When they sought the nomination, they have always been nominated. Humphrey... Um...
Moving on...
Bush the Elder in '88, Gore in '00.
The complication here is Biden. He didn't run in '16. Would he have lost to Clinton? Likely. This is a sort of selection effect, but let's just review the years we have left after the ones knocked out of contention by simple decisions, in the MF-er era.
Yeah, I like that term. I would, wouldn't I?
The '72 and '76 Democratic contests. The '80 Republican contest. The '84, '88 and '92 Democratic contests. The '96 and 2000 Republican contests. The '04 Democratic contest. 2008, for both parties. 2012 for the GOP, and 2016 for both parties.
That's 14 data points.
That's hard enough. We can do something with 14 data points. General election forecasting models in the post-WWII era developed at a time with similar limitations, back when good, ole' Ed Tufte got us started with this. Here's the thing, though. The basic candidate/party dynamic in a general election is always the same. D versus R, and maybe there's an "I" who will lose, but D versus R, and the economy matters.
Intra-party dynamics can change, and the candidate dynamics-- even the number of candidates-- can change far more in the nomination process.
A more fluid process doesn't mean science won't work. It just means we need more data.
Which we don't have.
That's why our models suck.
At this point, then, I refer back to how this blog started, with a series of posts called, "Trump to Political Science: Drop Dead." The references were twofold. First, there was a famous, old headline: "Ford to City: Drop Dead." 'Cuz I'm obscure. At the time, though, (early 2016), my fellow political scientists were confidently asserting that Trump had no chance at the GOP 2016 nomination, and they were doing so on the basis of political science models of presidential nomination politics.
Those models sucked then, and still suck now. My point at the time was that Trump was demonstrating how sucky the models were, and that we needed to understand why those models sucked rather than engage in rectal haberdashery.
Scientific models can be evaluated in two respects: their ability to make prospective forecasts, and their ability to pose retrospective explanations. Because we suck at explaining ourselves, we political scientists use the word, "predict," in both contexts. Because we are bad at everything. Also, our nomination contest models suck at both. We have neither the ability to predict prospectively nor explain. Limited data and heterogeneity in contest type.
I'll single out a few models here, as reminders. First, The Party Decides. This absolutely wretched, terrible book was always wrong, and always obviously wrong. A lot of people bought into it, and it got a lot of press, wrongly, by credulous fools during the 2016 contest because nobody actually studied the details of presidential election history and data. I always saw through this rubbish book. As a reminder, here was the argument. In the early days of the MF-er era, the unwashed masses controlled the nomination process and the muckety-mucks lost control, leading to people like... McGovern (the M of the MF-er), and Carter. Then, over time, the "party" muckety-mucks reasserted control over the nominations through endorsements, signaling, 'n stuff. Therefore, Trump couldn't be nominated because party muckety-mucks hated him.
And they absolutely did in 2015 and 2016. It is hard for people to remember this now, with the GOP having been thoroughly beaten into submission by this raging, psychopathic, authoritarian bully whose party has made a strategic calculation that they need to circle the wagons to protect its president from the consequences of his too-many-to-name scandals, but in 2015 and 2016, my fellow political scientists wrote off Trump completely because GOP officials hated him so much, and a worthless book called The Party Decides told them that this meant he couldn't possibly get the nomination.
I spent much of early 2016 on this blog calling bullshit on my field for their fealty to The Party Decides.
Now, let's say The Party Decides hadn't been debunked by 2016. What prediction would it make about the Democrats this year? Um... Uh... Maybe you see the problem. In principle, devotees of that ancient religion, which is no match for blasters, might say that, eventually party muckety-mucks will rally around someone and endorse that person, and everyone else will drop out, but around whom, and how does that work in a crazy field like this? Bluntly, if you are really a devotee of The Party Decides, then right now, you have to bet on Biden. Party muckety-mucks get nearly everyone else to drop out, rally around Biden, signal to the voters that he's the dude, and that's that.
Anyone predicting that? Keep in mind that the polls right now are all about name recognition. In principle, maybe they could rally around someone else, but around whom? This year? How does any of this work in the Democratic Party today?
OK, what else? I spent a bunch of time then, and still now, writing about a blustery model of wrongness by Matt Grossman & David Hopkins. The Republican Party is an ideological movement seeking purity, while the Democratic Party is a coalition of interest groups. That's why the GOP nominated their most ideologically pure candidate in 2016. Right? Right? Yeah, right. That's why a bunch of people got tax increases, we've got a trade war on our hands, and national security secrets are being handed over to fuckin' Russia in the Oval Office by a dude who lusts after his own daughter.
Here's the secret of the Republican Party. It isn't about ideology. It is about identity. White identity. Both parties are motivated by identity politics. One is just more homogeneous than the other. Philip Converse. "The Nature of Belief Systems in Mass Publics." 1964. People aren't ideologues. A plurality, though, think in terms of group identification. The GOP is about group identification. White identity politics. In 2016, they nominated the most aggressively white candidate. That's not about ideological purity because Matt Grossman & David Hopkins are wrong, and always were.
Anyway, though, what about the Democrats? I spent a bunch of time in 2016 bashing them for what their model predicted, wrongly, about the Republicans. What about the Democrats? Are the Democrats a coalition of interests?
Um... well, sort of, all parties are. That's basically the definition of a party, making that not a very insightful statement. Grossman & Hopkins were trying to use some cross-party comparison to explain asymmetric polarization, and I think they didn't succeed, but the statement that the Democratic Party is a coalition of interest groups is almost a tautology. So, does it make any predictions about who will win in 2020?
No. Tell me that, and I have no idea who wins.
What else did I cover in 2016? Spatial policy analysis? Here's the problem with that. In a two-candidate election, whoever is closest to the median wins. There's no solution when there are 500,000 candidates.
Money? Not especially predictive, except that candidates without it have to drop out, eventually.
What's going to happen? There will be some winnowing. Some of these people will drop out. Who? I have no clue.
Well, some. Seth Moulton will drop out, and hopefully go hide under a rock somewhere. I'd also be happy to say lots of bad stuff about Elizabeth Warren, Bernie Sanders and some others whom I particularly detest. However, as a political scientist, what do I think will happen?
I don't have a model here. Or at least, not a good one. Limited data, heterogeneity of contest type in the observations I have, and the models I have are crap.
So, what should you watch?
Watch the winnowing process. In the not-too-distant future, candidates will start dropping out. This will be a function of some combination of lack of money, lack of media attention, poor polling, etc. These factors are related to each other. However, candidates can get weird, little boomlets in the polls with media attention.
Is that what is happening with Pete Bu... oh, damn it. What's with people who have weird, hard to spell/pronounce names that start with "Bu?" If he doesn't go away soon, I'm going to have to figure this out. Anyway, how should I know? I'm a political scientist, not a...
Oh, right. That's supposed to be my job. We just suck at it.
Related Posts :
Michelle Obama Tells Oprah Winfrey America Lacks Hope And Will Miss 'Having A Grown-Up In The White House'
Former Chairman Of The U.S. Joints Chiefs of Staff: Russian Hackers Hit The Pentagon In A Crippling Attack In 2015
Venezuelans Face Cash Crisis As Replacement Bills Fail To Arrive
Hillary Clinton Blames Putin And The FBI Director For Her Election Loss- Finding humor in nuclear war, Walker's judgement plunges