No End of Ideology, Round 2

I am glad my piece on the netroots and ideology attracted the thoughtful comments of Mark Schmitt and the Prospect's own Ezra Klein and Matt Yglesias. In a companion response on my blog, I have answered the criticisms of bloggers and activists Chris Bowers and Sterling Newberry. I urge anyone who is interested in this debate to consult both responses. Answering the substantive criticism of Mark, Ezra, and Matt deserves a thoughtful response, and I hope I can narrow our differences here.

I do want to emphasize from the outset that the arguments I have made and make below are mine and not the views of my employer, The Democratic Strategist. The Strategist takes no stand on this or any other issue, being dedicated to empiricism and to engaging all factions of the Democratic Party.

Let me start with Ezra's comments. Ezra argues first that the netroots is so politically diverse that its members will inevitably throw their support to different Democratic candidates in the 2008 primary. This diversity will thus splinter a bloc that might otherwise -- if it were sufficiently out of the mainstream -- have enough power to harm Democratic prospects by propelling an overly liberal candidate to the nomination. But he also disagrees that the netroots is so ideological that they would ever back such a candidate.

I think Ezra's first point is largely correct. In the latest instant run-off poll (which isn't anything like a random sample of the netroots or even of MyDD readers, but which is instructive nonetheless), there are clearly four candidates with significant support: Wes Clark, with 30 percent of the vote in a full field of challengers, John Edwards (25 percent), Russ Feingold (20 percent) and Mark Warner (15 percent). In a field winnowed down to Clark, Edwards, and Feingold, each candidate still commands at least a quarter of the respondents' support, and in a head-to-head Clark/Edwards race, Clark has majority support, but only by a 53 to 47 percent margin. In other words, if the netroots' support in the primaries mirrors this survey, the group will be split down the middle between two mainstream candidates.

So how could the netroots hurt the party's electoral chances? And for that matter, if they unite behind Clark and Edwards, how can I claim that their liberalism will hurt the party? The answer to both objections is that the netroots can affect the positions candidates take, even if their support is initially dispersed among a number of candidates, and even if the candidates are popular with more moderate Democratic voters. John Kerry's vote against the Iraq supplemental spending bill is Exhibit A. In a four-way primary contest in 2008, Clark, Edwards, and Warner would compete for Feingold's supporters -- indeed they already are doing so -- and this could lead to a race to the left. It's easy to say now that the public supports the netroots position on Iraq. That wasn't true in the run-up to the war. Is it unreasonable to think that the netroots might pull the party away from swing voters in future primaries on an issue where they misinterpret the relevant politics? Given the evidence I cited in my piece indicating that potentially two thirds of the netroots could make such a misinterpretation, I would argue that it is not.

One of my disagreements with Mark highlights the dynamics that might exacerbate this potential problem. Mark argues that the netroots are more tolerant of moderate positions on issues on which interest groups have historically been unbending. I don't contest this claim. But while Mark thinks that the netroots' broad liberalism -- in contrast to the parochial liberalism of interest groups -- will benefit the Democratic Party, I agree with Noam Scheiber that ultimately there will be pressure from them for the party to move leftward in general. Consider how electoral pressure is resolved under interest-group liberalism. For simplification, assume there are exactly ten issues that voters are interested in. Interest groups of three equal sizes are organized around three different issues. One bloc demands a pro-choice candidate, another requires a protectionist trade policy, and the third calls for more domestic spending. Each of these groups is divided on issues other than its primary one or has weak preferences on them. The remaining voters -- a majority -- are divided or have weak preferences on all ten issues. A Democratic candidate can build a majority by toeing the line on just three issues (or fewer, if a plurality can be constructed from fewer than three of the interest groups).

Contrast this situation with a future netroots-led New Politics. This time assume one third of netroots voters will tolerate an anti-abortion candidate but requires orthodoxy on the other nine issues; one third will put up with a free-trader but no other stray positions; and one third will go along with a candidate's fiscal conservatism but will not bend on other priorities. The rest of the Democratic coalition is again divided or has weak preferences on all issues. Now a Democratic candidate must take the liberal position on all ten issues. In reality, the blocs of netroots voters are likely to be unequal in size and to be more pragmatic than in this setup, but the basic point stands that Democratic candidates must take a greater number of conventional liberal positions than in the interest-group-dominated world.

To oversimplify a bit, today there are only two issues the netroots demand conformity on -- opposition to an open-ended presence in Iraq, and support for grassroots populism -- and they are fairly united around these limited priorities. But what happens once Iraq recedes into the background? Eventually, the netroots might splinter, consistent with Ezra's argument. And that will tend to produce a situation closer to the ideal New Politics dynamics in my hypothetical example, with some splinters prioritizing cultural liberalism, others economic liberalism, others foreign policy liberalism. In this situation, even if everyone is fairly pragmatic, without the cohesion that an issue like Iraq brings the result is strong pressure from netroots members to toe the line on a broad array of issues.

Ezra argues that ideological factors don't explain the patterns of support and opposition displayed by the netroots. He says that the netroots oppose the Democratic Leadership Council (DLC) and The New Republic because those camps attacked them rather than because of ideological differences. But the DLC rift was arguably started by Dean's “Democratic wing of the Democratic Party” rhetoric and his criticism of “Clintonism,” both of which unfailingly drew applause from the netroots and embodied ideological differences portrayed as tactical differences. The New Republic dispute stemmed from Jonathan Chait's Dean-o-Phobe blog and the magazine's support for the Iraq invasion, the latter of which was clearly an ideological disagreement and the former of which was ultimately rooted in ideology as well (expressed as impressions of Dean's electability).

Ezra also says the netroots' support of the moderate Dianne Feinstein and NDN is a problem for my theory. But Feinstein doesn't make a New Democratic critique of the party. Nor is she much more moderate than Edwards, according to the respected Poole-Rosenthal scores. I'll take Ezra's word that the netroots support the NDN, but my own impression was that while Markos Moulitsas Zúniga supports the NDN, there's no particular love among the rank and file for it. Markos recently promoted NDN to his readers by contrasting the role of ideology between it and the DLC:

While affiliated with the DLC years ago, NDN long ago went its seperate way. [sic] In addition to its traditional PAC role of funding political candidates (where ideology plays a much smaller role than with the rigid DLC), the organization is becoming a model 527. [italics added]

So the distinction for Markos is either partly ideological or he feels the need to portray it as such to persuade his readers.

Mark questions why I would interpret admiration for Dean's willingness to take unpopular positions as an indicator of his “ultraliberalism,” but in my piece I actually interpret it as a sign that part of the netroots believes ideology is unimportant in elections. I'm sure that the netroots members who expressed such admiration, as Mark suggests, did so out of respect for Dean's authenticity, and that may have been a selling point for Dean among non-netroots voters. But the point remains that netroots members who thought unpopular positions would win enough voters looking for authenticity to replace the voters who would abandon Dean because of his ideology were endorsing the view that ideology ultimately matters little. Ironically, such views are evidence that the netroots' ideology -- across-the-board liberalism -- does enter into their decisions: If the content of issue positions doesn't affect election outcomes, then why not fully embrace one's liberalism?

Mark also speculates that the substantial agreement that John Kerry lost the election because he was too conservative may reflect respondents' belief that he was too cautious rather than too ideologically conservative. However, the survey wording for this option was, “Kerry's positions were too conservative,” and another option was, “Kerry's positions were too liberal” (which was endorsed by just 16 percent of the netroots, compared with the 41 percent saying he was too conservative).

Matt Yglesias rightly points out that it's not unreasonable for the netroots to believe that Kerry lost because his position on Iraq -- arguably the biggest single issue in the election -- was too conservative (because he should have opposed the war from the beginning). To address this question, I examined whether Dean activists who believed Kerry lost because he was too conservative were themselves more liberal than other Dean activists. I combined netroots members with other Dean activists in order to increase the numbers in my analyses, thereby increasing the precision of the results.

Dean activists who thought Kerry was too conservative were more liberal generally than other Dean activists -- 89 percent identified as liberal and 50 percent as very liberal, compared with 79 percent and 32 percent among other Dean activists. They were also over twice as likely to choose Dennis Kucinich, Carol Mosley Braun, or Al Sharpton as their second choice for president compared with other Dean activists (33 percent versus 14 percent when one looks just at those who voted for Dean). While hardly definitive, this evidence suggests that at least a sizeable fraction of the netroots members who thought Kerry lost partly because he was too conservative were not referring narrowly to Iraq.

Which brings me finally to Mark's objection that my data is inadequate for describing today's netroots -- or even that of 2004. I have always felt that bad data is no better than no data, and sometimes even worse. But good data is always better than no data. For describing the netroots, the Pew data I used is the best that is publicly available (contrary to Chris Bowers's claim). But is it good enough?

I don't think there's reason to worry that my analysis -- by relying on the Dean-affiliated Democracy for America database -- misrepresents the netroots of 2004. Through mid-2003, the netroots and Dean activists were nearly synonymous. In the second half of 2003, Wes Clark's campaign -- which came into being as a result of an online Draft Clark movement -- attracted significant netroots support as well. By the end of 2003, as the Iowa caucus loomed, a number of campaigns had large memberships, but Dean's dominated. Dean's 150,000 MeetUp members exceeded by fifty percent the number of all the other candidates' members combined.

Note that the Democracy for America database includes anyone who had participated in the Dean campaign in any way up to late 2004, so it includes individuals who went on to support other candidates and serve as online activists for them. It is worth comparing the votes for the Democratic nominee of the people in the DFA database to the number of MeetUp members affiliated with each campaign at the start of the primary elections:

By this rather crude comparison, the DFA data and my “netroots” sample include too many Kerry and Edwards supporters and not enough Kucinich and Clark supporters. Of course, this comparison is problematic in a number of ways, the most important of which is that Clark, Dean, Kucinich, and Edwards received fewer primary votes than they might have because each either dropped out or was effectively out of the race before some sample members were able to vote in their state's primary. By the same token, Kerry and Edwards received more primary votes than they might have if other candidates hadn't dropped out. But there is no getting around the fact that compared with the MeetUp figures, Clark supporters are underrepresented.

If we could assume that the actual netroots members who supported a candidate but were not in the DFA database were no different than the netroots members who supported the candidate and were in the DFA database, then we could weight individual responses depending on who they voted for in the primary so that the weighted support for each primary candidate matched his share of MeetUp members. When I did so, the only meaningful changes from data cited in my original piece were that only 30 percent (rather than 40 percent) said that Dean's willingness to take unpopular views was a draw for them, and just 54 percent (rather than two thirds) voiced either that sentiment or said that Kerry's conservatism hurt him in the election.

As for whether the netroots of 2004 resembles the netroots of 2006, there is frustratingly little evidence to go on. It is noteworthy that as recently as August of 2005, Chris Bowers of was referring to the Pew study as representative of the netroots and emphasizing its ideological nature:

Howard Dean remains the living, patron saint of the progressive netroots. As you can gather from this chart [from the Pew survey], netroots Democrats [respondents to the Pew survey] overwhelmingly self-identify as liberal or progressive, vehemently oppose the Iraq war, favor gay marriage, etc. Thus, while Democrats who take the structural and organizational challenges facing the party seriously will receive tremendous support online no matter their ideology, when it comes to choosing something like a presidential nominee or a committee leader, our first choice will be to go for someone who is both interested in reform and is a progressive. (Of course, since we are also "pragmatic," we do also support candidates who we believe can win. Thus, for example, we witnessed Feingold's precipitous fall from his original position as the 2008 netroots favorite early this year following his divorce…) [italics added]

I'm not sure I could sum up my argument about the netroots more succinctly. Note that “pragmatism” here refers to low support for Russ Feingold only in the wake of his divorce rather than out of the determination that his views make him unlikely to win.

Bowers claims elsewhere that the netroots stopped growing in September 2005. Assuming that this comes from reliable data that he has access to, the implication is that if the Pew survey represented the netroots in August 2005 -- as Chris indicated it did at the time -- it probably represents the netroots of August 2006. At any rate, Markos Moulitsas was already claiming around the time the Pew survey was conducted that the netroots wasn't ideological. He was wrong then, and I see no reason to believe that the situation is any different today.

Scott Winship is the managing editor for The Democratic Strategist, for which he also blogs, and a doctoral candidate in social policy at Harvard University. The views expressed do not represent the positions of either organization.

* * *

If you enjoyed this article, subscribe to The American Prospect here.

Support independent media with a tax-deductible donation here.