All “X Woulda Won” Arguments Are Essentially Useless

Jeremy Corbyn’s impressive performance in the snap election has led to another round of “Bernie woulda won” assertions. To state the obvious, this is a taunt, not an argument. A narrow defeat by the out party in a completely different political and media context says absolutely nothing about what tactics or candidates would have won the 2016 American presidential election. And nobody really believes that it does.  Is anyone making “Bernie woulda won” arguments going to argue that Macron — who actually won the presidency and whose new party won crushing legislative majorities — is evidence that neoliberalism is the best weapon against fascism and the Democrats should emulate him? Of course not. And they’re right! But this is because of substantive policy reasons, not political reasons. And I wish people would just say that, rather that this bad faith Mark Halperin crap about how my policy preferences are always the best politics in every election in every jurisdiction.

Well, almost nobody. Ygelsias tries to salvage the argument by taking it out of the realm of the pundit’s fallacy:

This isn’t either faction’s favorite frame but I think “Bernie would have won” follows from “Comey & Russia cost Clinton the election.”

— Matthew Yglesias (@mattyglesias) June 9, 2017

Yes. I think there’s a problematic ambiguity in BWHW rhetoric even though I agree with the claim.

— Matthew Yglesias (@mattyglesias) June 9, 2017

I think this is entirely possible, particularly with a generic white guy Dem like Biden or O’Malley. Sanders I dunno — I have a hard time imagining political reporters for whom deficit and entitlement cuts are some of the view things where Views Do Not Reasonably Differ warming up to Bernie, but then compared to Clinton media treatment is pretty much all upside. And presumably the Director of the FBI would not have baselessly implied than any of them were crooks less than 2 weeks before the election. And if Bernie might have got less favorable media treatment than a generic candidate he also has base mobilizing potential that they didn’t.

Having said that, this line of reasoning still doesn’t strike me as very useful:

• The logic here — a candidate with higher favorability ratings would have done better — is solid. But nobody knows how another candidate’s favorability would have held up during a campaign. Hillary Clinton was popular before she started running for president too.
• The argument has an implied ceteris paribus condition. The election came down to less that 100,000 votes. Take away the director of the FBI baselessly implying that Clinton was a crook and generating a massive wave of negative coverage less than two weeks before the election, and the Democratic candidate almost certainly wins, all else being equal. This is true as far as it goes.  But you can’t just assume a Democratic candidate is ahead on October 27.  This suggests that there’s nothing that Clinton did well that a generic candidate wouldn’t, and I don’t think that’s true. In particular, she’s the best debater of any major Democratic candidate of my lifetime — Obama very much included — and her pasting Trump in the three debates and winning the media spin on them set her up to win had Comey not decided to go Jim McAllister. A lot of people seem to think that beating Trump in the debates was a no-brainer, but (and ask Rubio, Jeb!, Cruz et al.) that’s really not true. As anyone who’s done it knows, debating a bullshit artist/insult comic is not easy, at all. I think Bernie and especially Biden might not have shown her discipline and preparation. O’Malley who knows, but…sorry, I just nodded off trying to remember him trying to speak. But at any rate, you can’t just use an alternative candidate to change one thing about the campaign — a different candidate has their own mix of strengths and weaknesses, and in a structurally close election you can’t make an assumption that holds everything but one event constant.
• Still, given Clinton’s high negatives I don’t think the first two points are the biggest problem with the counterfactual. The real problem, as I’ve mentioned before, is that when you posit a Sanders (or O’Malley or Biden or Webb) who can beat Clinton, you’re positing a different and non-existent candidate.  A Bernie Sanders had spent decades cultivating relationships with major Democratic constituencies and done well enough to hold his own in Southern primaries would be a very strong candidate — but Bernie didn’t actually do that. Speculating about how Martin O’Malley would have done had he won the Democratic primaries rather than getting roughly six votes strikes me as about as useful as speculating about how Wonder Woman or Abraham Lincoln, Vampire-Killer would have done in the general. Sure, an O’Malley who could win the nomination would probably be a pretty strong candidate, but he also doesn’t exist. A Joe Biden who was committed to running, disciplined, persuaded Obama that he was the better candidate, and didn’t have two zero-delegate primary runs under his belt…you get the idea. It’s OK to discuss alternatives as a parlor-game hypothetical, but it’s really not a very useful exercise either retrospectively or prospectively.

One error people are making in analyzing the British elections is conflating beating structural limitations with beating expectations. In the abstract, it’s not all that surprising that Labour would pick up ground in a snap election called by a weak leader of a party that had just committed a massive political blunder. The outcome was surprising because a lot of people thought Corbyn was an even worse candidate than May. Granting that for a while polls seemed to reflect this, the conventional “wisdom” turned out to be completely wrong.  But if the lesson you take from this election is that we should be very confident about who would or wouldn’t be a good candidate ex ante, I would suggest you reconsider.