It's important to win Presidential elections and not lose them. So it makes sense for Democrats to care about which primary candidates are more likely to win than others. Unfortunately, Democratic thinking on electability has followed a bad strategy: choosing the candidate with the most salient Republican cultural signifiers.
Today those are the whiteness and maleness of Joe Biden. But in late 2003, deep in the shadow of 9/11, it was John Kerry's record as a decorated war veteran. Unfortunately, a Republican smear campaign cast doubt on Kerry's reputation (some will recall the Swift Boat Veterans for Truth). Because of that and other things, Bush was re-elected in the depths of the Iraq War.
In 2007, still in 9/11's long shadow, probably nobody theorized that Democrats had to nominate an African-American whose dad was from Kenya, whose middle name was "Hussein", and whose surname rhymed with "Osama". But we nominated him, and he won the election and the next election.
The point here is that having Republican cultural signifiers just doesn't predict that much about how likely you are to win the general election. Maybe it helps a little in winning over persuadable Republican voters, which I guess is the whole idea. But that effect is mediated by big random things and can't be relied on.
Some people say that you should just vote for whoever you prefer, and that'll guide you to the most electable candidate. That's probably better than the Republican cultural signifier method. But we want to go beyond our own idiosyncracies, and I think we can do that.
My main method is to look at polling data. In particular, I like looking at candidates' national favorability ratings in a good selection of national polls, right when it's time to vote. In early 2008, Obama was pretty consistently getting net favorability numbers over +20 and occasionally over +30 (the best was 61 favorable, 27 unfavorable). This doesn't predict everything. The campaign can go badly, as it did for Kerry. But if a candidate is scoring high without exhibiting the Republican cultural signifiers, it's a sign of political talent, which Obama definitely had.
(Democrats also seem to do better when nominating relative neophytes like Obama and Bill Clinton than when nominating old party hands like Mondale, Dukakis, Gore, Kerry, and Hillary Clinton. Maybe a shorter record just has less bad stuff. I count inexperience as a slight plus for electability, which is weird but fits the data too well. Maybe the favorability thing encompasses the whole effect? I want more data.)
Anyway, don't worry too much about electability now. There's a big campaign ahead of us. There's plenty of time for Elizabeth Warren to build a national reputation for having awesome ideas about how to give you more money, and put that weird Native American scandal behind her.
But if you want to think about it when the time comes, look at polls. They're a survey of the voters, and the voters decide. The information you want is there.
Today those are the whiteness and maleness of Joe Biden. But in late 2003, deep in the shadow of 9/11, it was John Kerry's record as a decorated war veteran. Unfortunately, a Republican smear campaign cast doubt on Kerry's reputation (some will recall the Swift Boat Veterans for Truth). Because of that and other things, Bush was re-elected in the depths of the Iraq War.
In 2007, still in 9/11's long shadow, probably nobody theorized that Democrats had to nominate an African-American whose dad was from Kenya, whose middle name was "Hussein", and whose surname rhymed with "Osama". But we nominated him, and he won the election and the next election.
The point here is that having Republican cultural signifiers just doesn't predict that much about how likely you are to win the general election. Maybe it helps a little in winning over persuadable Republican voters, which I guess is the whole idea. But that effect is mediated by big random things and can't be relied on.
Some people say that you should just vote for whoever you prefer, and that'll guide you to the most electable candidate. That's probably better than the Republican cultural signifier method. But we want to go beyond our own idiosyncracies, and I think we can do that.
My main method is to look at polling data. In particular, I like looking at candidates' national favorability ratings in a good selection of national polls, right when it's time to vote. In early 2008, Obama was pretty consistently getting net favorability numbers over +20 and occasionally over +30 (the best was 61 favorable, 27 unfavorable). This doesn't predict everything. The campaign can go badly, as it did for Kerry. But if a candidate is scoring high without exhibiting the Republican cultural signifiers, it's a sign of political talent, which Obama definitely had.
(Democrats also seem to do better when nominating relative neophytes like Obama and Bill Clinton than when nominating old party hands like Mondale, Dukakis, Gore, Kerry, and Hillary Clinton. Maybe a shorter record just has less bad stuff. I count inexperience as a slight plus for electability, which is weird but fits the data too well. Maybe the favorability thing encompasses the whole effect? I want more data.)
Anyway, don't worry too much about electability now. There's a big campaign ahead of us. There's plenty of time for Elizabeth Warren to build a national reputation for having awesome ideas about how to give you more money, and put that weird Native American scandal behind her.
But if you want to think about it when the time comes, look at polls. They're a survey of the voters, and the voters decide. The information you want is there.
