Pages

Thursday, July 19, 2012

Who's getting primaried, and why?

Hans wrote two interesting posts recently (here and here) on the subject of asymmetric polarization. That is, he wondered why DINOs (Democrats in name only) aren't singled out for retribution by liberals the way that RINOs (their GOP equivalent) are by conservatives. I share Hans' impression that there's a difference across the parties here, and that this almost certainly has something to do with the fact that Republican officeholders seem to be moving right far more quickly than Democratic officerholders are moving leftward.

I wanted to point out, however, the work of Robert Boatright, who has a book coming out on congressional primaries. As Boatright notes, contrary to much speculation, primaries aren't becoming more common in congressional races, and ideology is only one of many motivations for primarying an incumbent. Primaries, he notes, are most common within a party that is having a good year; the same things that motivated Republicans to challenge Democrats in 2010 also motivated Republicans to challenge sitting Republicans.

Perhaps most interestingly, he finds some important differences between the parties in terms of ideological primary challenges, and not in the expected direction. Below is a scatterplot of Republican House members from 1972 to 2008, plotted by the ideological makeup of their districts and their ideal point on congressional roll calls. Those who were challenged in primaries are highlighted. The important thing to note is that some incumbents who were primaried were too liberal for their districts, but some were too conservative:

Here's the same graph for Democratic House members. Interestingly, basically every primaried Democratic incumbent was too conservative for her district. This makes it look like Democrats are more the party of ideological purity than Republicans are.
Now, this dataset encompasses the 1970s and ends in 2008. It is quite possible that the roles have reversed in recent years, with Republicans embracing ideological purity and Democrats becoming more tolerant of moderate behavior by incumbents. But there seems to be nothing endemic to the Republican Party that would make them more obsessed with the purity of their officeholders.

6 comments:

  1. What I like most about this is the application of real data. My impression is that conservatives are holding their members accountable more. The conventional wisdom is that this is in part due to primary-ing, but the Bailey, Mummolo and Noel paper on the Tea Party is not actually about primary-ing, but other sources of influence. So maybe the CW is wrong about how they hold members accountable.

    But that also means I could be wrong that they are better at holding people accountable. I think they are not, because of where the parties are.

    Part of the issue is what constitutes "too moderate" for the district, since presidential vote share and nominate scores don't have a natural translation. There's just the average. If the GOP is farther to the right, they might be on both sides of the true translation, while Dems may not be.

    ReplyDelete
  2. A couple questions, beyond the one Seth raises about the possibility of change over time.

    One is that I sure would like to see the South separated out (for the Dems of course).

    The other is that we're not seeing here which direction the challenges are coming from. So it could be that all the GOP challenges are from the right, even in districts where the Member is too conservative for the district. So what it *could* be showing is that Dems use primary challenges to move the House to the left, but Republicans use primary challenges to move the party to the right. (It's also possible that some of these challenges are coming from the "wrong" direction anyway even when we wouldn't expect it, such as a Dem challenge from the right even though the Dem Member is too conservative for the district.

    Obviously that's all speculative, and I agree with Hans that it's good to have some real evidence, but I'd be careful about overinterpreting just what we have here.

    ReplyDelete
  3. This question of the "wrong" direction is kind of what I'm getting at with the no "natural translation." Suppose, following Masket and Noel 2011 (how do I put links in comments?), that elected officials are ALWAYS more extreme than their district. Then it's not possible to be primaries because you are too moderate FOR YOUR DISTRICT, because you never are. But what can happen is that you are too moderate for where the party (base) wants you to be, but your party will give you some leeway for where your district is.

    So then the question is how far can the party/ideologues pull members? If one party increases how far they are pulling, then they will appear to be primary-ing people who are already "too extreme."

    ReplyDelete
    Replies
    1. How many of the challengers who won the primary held the district in the next cycle?

      Delete
  4. 1) I just can't believe those data. That's amazing. I'm going to have to read that book when it comes out.

    2) At some point, the presidential vote share measure begins to break down, and doesn't capture Hans' point: how far left or right is the party base in that district? For that matter, with national fundraising now perfectly common for House members (particularly the ones given to more crazy pronouncements), do we need to think about TWO primary bases: the partisans of CD_XX and the national base? (Which, I think, is getting at Hans' point, though not quite)

    ReplyDelete
  5. This is partially a methods issue. Y'all are right that presidential vote has its limits as an indicator of district ideology and doesn't map naturally onto NOMINATE scores. Some sort of method like the one Bafumi and Herron used or Hans and I used would probably be helpful here. But also Jarv is right that there are multiple conceptions of "constituency" and it's not obvious which is the right one and which is more deadly to be out-of-step from.

    ReplyDelete

Note: Only a member of this blog may post a comment.