Hans Noel and Seth Masket discuss the pitfalls of electability arguments, and what might make for good evidence.
We know there are problems with electability arguments, but to what extent should party elites still be thinking about it? We decided to sit down and discuss it.
Hans Noel: You just wrote a piece arguing that electability isn’t a very useful concept in choosing presidential nominees, mainly because it is used as a proxy for ruling out women and underrepresented minorities. The piece was pretty nuanced, but the takeaway, from the headline to how you and others have discussed it on social media, has been that we should just abandon “electability.” Are you really saying that a party shouldn’t have a discussion about whether its candidate can win in the general election?
Seth Masket: We could have a whole conversation about how to write a nuanced headline — but my take is not that people shouldn’t discuss or think about electability at all. Rather, it’s that the concept of electability is a) not well defined, b) not well supported by evidence, and c) conflated with stereotypes about female candidates and candidates of color. So we should be extremely cautious when making this kind of argument.
HN: I suppose not supported by evidence depends on what definition we’re using. The term “electability” is confusing. It sounds like you either have it or you don’t. But broadly speaking, I think it’s coherent to talk about candidate-specific characteristics that affect election outcomes. And there are a lot of them.
Some of them we have some evidence on, no?
HN: And some evidence that scandals hurt, although that could be because most allies and donors abandon the candidate.
SM: Right. And there’s some evidence that white voters are less supportive of African American candidates, although it’s possibly because they think the African American candidates are too liberal. So maybe that’s race, maybe that’s ideology, or maybe they’re hopelessly conflated.
HN: But these effects are all pretty small, and maybe especially small in presidential elections.
SM: Exactly. If 2016 showed anything, it’s that it’s pretty hard to get Democrats to vote for a Republican and vice versa, even if your own party’s nominee is quite unpopular.
HN: So I think we’re starting from the same place. When you think about the most important things to understand about a presidential election, candidate quality is not at the top of the list. Is the incumbent running? What is the state of the economy? Are we involved in a foreign war? What is the partisan division in the country? You can answer all those without knowing the candidates’ names. But I think part of that is that the nomination process, for all its flaws, usually picks candidates who are pretty appealing. If a party actually nominated Jonah Ryan, we’d see what that looked like.
SM: Sure, but arguably, 2016 provided that. Donald Trump kind of went out of his way to be unappealing, and 90 percent of Republican voters convinced themselves that he was still better than the Democrat.
HN: True. Although, Trump underperformed the fundamentals. If 80,000 people in four states had voted differently, we’d be saying that Trump proves candidate quality matters. What should have been an easy GOP win was actually incredibly difficult and maybe even a little lucky.
Meanwhile, one argument we hear is that Hillary Clinton could have won in 2016 if she had the mobilization of black voters that Obama had. That’s candidate quality too, although I don’t think anyone would say Clinton was “unelectable.”
SM: That’s fair. I suppose my main concern is the way “electability” concerns are used during the nomination process. I’ve seen and heard a number of arguments that only a white male Democratic presidential nominee can beat Trump. The evidence doesn’t really show that. But it’s apparently a pretty compelling argument for many, and it can be hard for candidates to overcome that perception.
HN: I’m in agreement with you here. There’s a case to be made that a woman or candidate of color has an advantage in the general election, because they would mobilize voters that a white dude can’t mobilize. If black voters had voted in 2016 like they did in 2008, they would have tipped Michigan and Wisconsin. But it’s not surprising that they were less excited about Clinton than they were about Obama. So race and gender should be part of the conversation.
SM: This is tricky, though. I’ve been leaning toward, “Let’s try to avoid the ‘electability’ argument since it hurts women and POCs,” and you seem to be suggesting, “No, let’s talk about it, but women and POCs may be more electable than white guys.” Is this right?
HN: At least partly, yes. The one thing that a wide-open nomination process does that also helps in the general election is that it can help nominate a candidate who is broadly acceptable to the party, which in turn mobilizes all the party in the general election. And I think it’s time that we think seriously about the Democratic Party, anyway, as being a party in which women and people of color, and maybe especially black women, have played a huge part. If you want to excite Democrats, maybe you think about what those parts of the party want.
SM: That’s a good message. But at least so far, the conversation has largely worked against women and people of color. That is, not only are many party elites convinced that white guys are more electable, but that information gets through to prospective candidates, and some women and people of color choose not to run as a result, knowing they’ll get less support. If African Americans (especially African American women) are more moving into the ranks of party elites — and they are in some parts of the country — that conversation changes. But not overnight.
HN: I agree. What’s interesting is that on race, we resist the changes. But on ideology, many people think that Bernie Sanders, who has a reputation as the most ideologically extreme person in the race, has no problem. The party has moved to the left, so it’s time.
But I think Sanders has an electability problem, not just because he’s ideologically extreme, but because he’s alienated a big part of the Democratic constituency. So he maybe can pick up some working-class white Trump voters, but he’s not going to do better than Clinton did with women and people of color.
SM: I think this is right. People employ the electability idea rather selectively. We have pretty good evidence that being an ideologue hurts, and less good evidence that downscale whites will vote for a white guy. I tend to think that Republicans will try to smear any Democratic nominee as a socialist, but it might be a more convincing argument if they can produce video of the Democratic nominee himself making that claim.
HN: I’m reminded of Matt Yglesias’s quip that “What Democrats need to do to win is enact my personal policy preferences.” That kind of myopic view is a problem. And someone who really wants to nominate a woman, an underrepresented minority, or a socialist could all be charged with it. But as you noted originally, we only interrogate the electability when it comes to women and POC.
SM: A couple of additional frustrating things about electability. One is that there’s an awful lot of faith placed in early primary polling as a predictor of general elections. People point to Sanders’s favorables in 2016 and say he would have been a much more popular nominee than Clinton, but Sanders never had to endure a major negative campaign from the GOP. He might have been less popular by November than she was. Similarly, we assume that just because Biden has high positives now, he will 18 months from now. We really don’t know.
Second, outside large-N academic studies, this is really hard to prove. Let’s say I somehow convince “the Democratic Party” that white men aren’t more electable. And they go ahead and nominate Kamala Harris, and she loses narrowly to Donald Trump. Then the party comes back and says, “You told us a black woman could win!” And I’d say, “No, I just said she’d do no worse than a white man.” And they’d say, “But Biden would have won!” And I’d say, “You can’t prove that.” And then we’re right back where we started.
HN: That’s true. That’s the same reasoning that concludes that Trump could win, even with all his baggage (so electability must not be about scandal) but Clinton could not (so electability must be about gender).
But what happens if we just stop talking about electability? Then we end up relying even more on early polls and primary outcomes. Whoever wins early elections, in unrepresentative states with tiny electorates and against different choices, is presumed to be electable. Maybe that’s the answer. The party should just step aside. But I don’t think so. I think primaries aren’t a very good way to figure it out, because the choice set and the constituencies are so different. Especially with the strange sequential system we have now.
SM: So this is an important question. If we allow that some discussion of electability is important, what is actually a good test? Like you allude to, there were some party elites in 2008 who were hesitant to back Obama until after he won the Iowa caucuses; they took that win as a sign that he could translate enthusiastic support into actual votes (unlike, say, Howard Dean). You’re right that it’s not really like a general election win, but is it irrelevant?
HN: No. In that case, it’s a lot like Kennedy winning in West Virginia in 1960. He showed he could win voters in a Protestant state. The problem is that Iowa and New Hampshire are super white, so they’re especially hard for candidates of color. So someone needs to interpret the results. Such interpretation is, in the end, a conversation about electability.
SM: Ah, you’ve been reading my book (which comes out next year). But yes, the interpretation of the last election is really important for how people assess electability!
For some, 2016 showed that a woman can do just as well as a man. (Clinton overperformed economic models and beat Trump by 3 million votes.) For others, her failure to win the Electoral College shows that women just can’t win.
HN: No (I mean, yes, your book is great, but), what I mean is, in the moment, people needed to interpret Obama’s win in Iowa as a signal that he could win elsewhere, but a narrow Obama loss in Iowa wouldn’t have meant he couldn’t win the general. If we don’t interpret the results during the process, we don’t get that nuance.
SM: Okay. Which means the people in charge of the interpretation are really powerful in determining the course of the nomination, right? So who does this? Is this a media role? Party elites? Bloggers? The Twitterati?
HN: Well, one argument is that’s it’s party elites. But they clearly don’t have a monopoly and maybe don’t have the influence they once did.
In our textbook (in progress), we say that the party has two tasks in choosing a nominee. They want the “best” candidate, defined in terms of how they would govern, who can also win, defined in terms of the general election. I think that balance is important and necessary. And I think “the party,” broadly constructed, should be thinking about it.
SM: We should probably wrap up. But here’s an analogy I’ve been thinking about. Electability is a lot like student numerical evaluations of professors. We know those scores are heavily biased against women, scholars of color, the LGBTQ community, and other underrepresented groups in academia. So I’d be tempted to just to ignore them altogether.
But there’s also some useful information in there. If a professor has been pulling all 5s in one class and then drops to a 3 one year, I’d want to know why. Did she try a new teaching method? Is she having some difficulties? Did she just draw a hostile set of students? Are there things we can do to intervene?
It strikes me as difficult, but not impossible, to separate out the useful information from the biases. It may be harder in presidential nominations. (Professors aren’t directly running against each other, generally speaking.)
HN: That’s a good analogy. And it’s maybe even harder because we don’t even have a plausible metric for electability that one could get a 5 or a 3 on, before the election itself, when it is too late. In the professor case, I think we’d recommend using the scores as a warning, and then looking more closely, but not using them directly as inputs into tenure, promotions, or raises. In the candidate case, I think we might say that there is no real way to get biases out of a political discussion. Which is a shocker, I know.
SM: I suppose we’re back at the point where we want people (party elites, in this case) to be making judgments using empirical evidence and trying to weed out bias, but I don’t know why they should be different from anyone else in any other field.
HN: They probably aren’t any different. But maybe in most cases, awareness of the bias, and especially increasing the diversity of those making the judgments, can go a long way.
Powered by WPeMatico