Why Is Hillary Running Left?

[Reginald]

It's his article that he links in the RCP article.

Quoting from the linked article:

You must be logged in to read this quote.

Bolding the most problematic section for me. Why are we assuming these are the same voters here? Those statistics for both the white and the Hispanic electorate are more Republican than in 2012, especially for the latter. Trende seems to assume that difference is mostly driven by individual vacillation, but it could just as easily be driven by a differential effect on turnout rates with regard to ideology/affiliation, i.e. perhaps the left-leaners within each ethnic group did indeed sit out 2014 moreso than those on the right. You can't really get at this by comparing relative shares of the electorate. Some of the comments on Enten's 538 article discuss this.

Trende does try his hand at messing with ideology:

You must be logged in to read this quote.

"Just" three points. So their margin would change from 5.7 points to 2.7 points in the House. And that's at the pretty useless national scale, so who knows what that would look like at the state or CD level. (Not to mention the problems with self-identification of ideology...)

Now obviously demographics don't tell the whole story re: the Democrats' "midterm penalty." I think the difference in Obama's approval ratings certainly holds a lot weight here. But this just seems like quite a declarative conclusion given the flaws in the analysis.

[Reginald]

Reginald, you seem to be looking for far more detail than one should expect from a pundit of any political stripe. Recombining everything at the level of the CD will probably turn up in a graduate student's thesis before too long. Look at the work and time Griffin is putting into his county-level analysis of white Dem support in 2012. A professional political analyst writing for a website isn't paid to take that amount of time, so until there is a research paper one is limited to nationwide analyses. That said, I have to add that Trende does a much better job of using real data to model across elections than most pundits.

I totally agree. There are a lot of moving parts here, and an op-ed or whatever isn't likely (or even liable) to capture everything that's relevant. But that's pretty much my point. It's quite the weighty conclusion given such insufficient analysis, and it should thus be taken with some skepticism.

More broadly though, the point about the future of the Obama minority coalition is apt. However, to that point, I'd submit that there might simultaneously be something of an anti-Obama coalition whose future needs to be kept in mind as well.