I’m not sure what to make of this analysis from my Washington Monthly colleague David Atkins. He’s taken on a task that is probably too ambitious, which is trying to predict how the Democratic primaries will go and what kinds of factors will drive them. I’d probably be dissatisfied with any blog-length effort to do this at such an early point in the process, including any I might offer myself.

Having said that, I do have specific critiques of the arguments Atkins has put forth. To begin with, he insists Elizabeth Warren and Bernie Sanders should be considered odds-on favorites because the Democratic socialist wing of the party is “ascendant,” but it’s not really clear what he means by that term. The socialists are certainly on the rise, but that does not mean that they are dominant. Some of their ideas are moving from the fringe of the party toward the mainstream, but that can cut both ways. The ideas are getting more support in part by being coopted or accepted, which means less of an advantage for those who pushed them first. Anyone voting on the basis of Medicare-for-All, for example, will have perhaps more than a dozen options to choose from this time around.

Then there’s the way Atkins categorizes the candidates. He’s putting Sanders, Warren and Gabbard in one box and everyone else in another. He categorizes Booker, Harris and Castro as “establishment” candidates, which must be news to them. For some reason, Sherrod Brown doesn’t even warrant a mention. He doesn’t really explain his reasoning for putting Warren in a camp with Sanders and excluding everyone else but Gabbard.

While he correctly notes that early polling is largely about name recognition, he doesn’t address the latest poll in the field out of Iowa which shows former vice-president Joe Biden still in a commanding position with 29 percent of the vote. That poll, like the more general polling average, shows Warren and Sanders collectively pulling about a quarter of the Democratic electorate. Both of them individually trail Kamala Harris in Iowa who is clearly riding a post-rollout surge.

In an effort to rebut charges that the socialist wing lacks support from African-Americans, he makes distinctions about age, noting that Sanders did better with younger blacks in 2016. He also insists without providing supporting evidence that people of color are not any more ideologically moderate than liberal whites. That’s probably wishful thinking in the one area that really matters for this analysis, which is understanding the basis on which people cast their votes. I doubt very much that a significant fraction of blacks will cast Booker and particularly Harris into a category of retread apologists for the status quo and establishmentarian policies. I also doubt that the residual love for Barack Obama in the black community will fail to lift Joe Biden’s support with them above what it might be simply based on some ideological test. Black voters are more pragmatic than ideological and they don’t vote strictly based on identity. That is why so many of them waited until they saw Obama win in the nearly all-White state of Iowa before getting on his bandwagon back in 2008. If Booker or Harris have early success, then it’s possible that black support for them will surge, but again without much regard for ideology.

There will be another group of Democrats strongly interested in having a woman as the nominee, and they will likewise rally to a woman who takes an early lead without a whole lot of consideration for ideology. Protestations aside, identity politics on the left is at least as ascendant as socialism, which is one reason Warren could win the nomination even though she’s been typecast (yes, even by Atkins) as some kind of radical choice. The problem here is that Atkins is arguing that she could win because of her ideology rather than in spite of it.

In an effort to argue from authority, Atkins sets out his credentials at the outset.

I was one of the few writers to consistently predict from early on that Donald Trump (or perhaps Ted Cruz) would win the GOP nomination over his more establishment foes (and that he had a very good chance of winning the general election as well.) It wasn’t just a hunch: it was based on a close reading of the GOP base as well as the basic polling.

I also predicted early on that Trump stood a very good chance of winning the nomination, although I never took his general election chances seriously enough. Where I’ve had predictive success in the primaries, going all the way back to 2004, is by looking very closely at the process of how delegates are actually awarded. I am unwilling to make predictions for 2020 until I have completed a thorough analysis of how that process will work this time around, as well as at the calendar.

More than that, though, I think the candidates need to be heard from before we begin categorizing them. I see enormous distinctions between Warren and Sanders that Atkins doesn’t acknowledge. I think Cory Booker is surprisingly strong on market consolidation, particularly in the agricultural economy. Kamala Harris supports essentially doing away with the private health insurance industry which really ought not place her in the camp that doesn’t threaten Wall Street. Sherrod Brown is bringing his “dignity of work” message which aspires to bridge the artificial gap between progressives and the labor movement. I don’t think the candidates have even begun to ideologically sort themselves, and many of them haven’t even finished developing their message or strategies. There will no doubt be a few people who run to the middle arguing that everyone else is dangerously radical, but they won’t get any traction.

Above all, with so many candidates and without winner-take-all contests, it’s going to be difficult for anyone to take a majority of the delegates to the convention. For now, I’d rather be Joe Biden than any of the other candidates. He’s in the lead and he can count on Obama’s neutrality if not his outright support. He has the most appeal to the widest ideological range and would be the logical consensus choice in any brokered convention. He has connections and media approval that no one else can match. And, more than anything, there are so many options that it will be hard for anyone to consistently beat him from contest to contest.

He may not be a logical fit for the times or even the mood of the Democratic electorate, but he’s in the strongest position. The biggest threat to him is an early winnowing of the field, but that seems unlikely to happen this time around. For one thing, on the delegate math alone, no one will be able to get much of an early lead. For another, Trump showed how you can run a primary campaign on a shoestring budget, and there are always billionaires who can keep people alive with Super PAC money and the power of small donations for any candidate who can get a passionate following.

While I am not making any strong predictions here, I also think Atkins underestimates how badly 2016 damaged Bernie Sanders’s standing with the Democratic electorate. He’s treating the quarter of the electorate that currently supports either Sanders or Warren as a moveable block, as if Sanders could command it were Warren to drop out. It’s more likely that the 15 percent Sanders is currently pulling in Iowa is close to his ceiling. He’d have to secure the nomination before hoping to win the support of a single Clinton voter from 2016, and his base of support is already splintered among other alternatives. I don’t see how he could win more than a small fraction of it back.

Having said all of this, I do think it’s likely that the eventual nominee will emerge with an ideological profile far to the left of what has historically been the norm. But I don’t think that it follows from that they will have necessarily started from that point on the spectrum. The Democrats are going to push the candidates to the left but they’re also going to be on the lookout for weaknesses that they’ll worry can be too easily exploited. The desperation to beat the Republicans in 2020 will be great enough that pragmatism will still play a gigantic role in how people vote.