ACA and 2010, Revisited

Lots of blogging talk yesterday about a new paper forthcoming from Brendan Nyhan, Eric McGhee, John Sides, Seth Masket and Steven Greene showing the effects of ACA on the 2010 elections. Their finding is that (1) the final ACA House vote seems to account for the difference between expected Democratic losses from the economy and other fundamental factors alone and the actual losses the Democrats experienced in 2010; and, (2) that the way that actually happened was that voting for ACA made a Member appear far more liberal, and therefore voters seeking to vote for the candidate ideologically closest to them flipped to the Republican. Seth explains it here; John explains it here, Brendan here, and here's the ungated paper. In my view it's an excellent contribution.

That said: I still have the same critique of it that I had when the preliminary studies were published, and which Kevin Collins tweeted, and which Jonathan Chait wrote about yesterday.

The two main points that Chait covers are: suppose that the bill failed in the House, with all vulnerable Democrats voting against it and therefore dooming it (but avoiding the effect that Nyhan et al. find). What's the effect to Barack Obama and to Democrats in general for failing to pass anything? (This one is complicated, as Chait points out, because the House had already backed the bill the first time around). That's a good question, and one which the authors themselves raise in their concluding discussion.

Or, what happens if the Democrats simply never move forward on the bill in the first place? Would some other issue simply have substituted for health care reform in GOP advertising, and thus have achieved the same effect -- and meanwhile, what damage (if any) would it have caused among liberals?

The truth is that we don't know. It's certainly possible that there was something special about health care reform that sparked such a strong reaction, and that (in the Nyhan et al. paper) convinced swing voters that Members were more liberal than they otherwise would have believed. On the other hand, it's also possible that with health care out of the picture that something else would have worked just as well.

Thinking about it...on the one hand, there's the evidence that those Members who didn't vote for ACA did not, in fact, get a substitute effect. That is, one would think that in those districts their opponents would simply plug in the stimulus, or cap-and-trade, or some provision of an appropriations bill and pound in the message that the vote for whatever it was proved that Member Smith was very liberal. Perhaps the GOP and their various candidates weren't flexible enough to do that. Or, perhaps the local advertising only worked when it was accompanied by the general  information environment we actually had. That is: no health care bill at all, and Rush and Beck and the rest pound on Substitute Issue B and therefore local advertising against Member Smith for voting Yes on B works. Given health care, and Rush and Beck and the rest bounding on it, then a lone campaign claiming that Yes on Substitute Issue B proves extreme liberalism doesn't work. Or, again, perhaps there's something particular about health care in general or the ACA in particular that's doing the work, and it's really true that without a bill, there's no effect.

That's pretty much the response that John gives this morning -- that for the Republicans, finding an effective Substitute Issue B is easier said than done. For example, on the stimulus John says "Critiques of the stimulus amounted to garden-variety “wasteful government spending.”  I’m not sure that’s as, um, stimulating as “National Socialist Health Care: The Final Solution” and “death panels” and the like."

But that raises another question. Even if it really is ACA that's doing the work, we still don't know what it is about ACA that's doing it. The mandate? The "government takeover"? The "death panels"? The Medicare cuts? The spending? The taxes? The notion that health care should be a "right"? Simply an association with Obama? Some of those would seem to sort with "liberal" more than others, but which ones were actually moving voters towards concluding that their ACA-supporting Members were very liberal? Of course, you can't do everything in one paper, so I hardly would blame the authors here for not answering that question. But without knowing what it was about ACA that was so offensive to these voters, it's very hard to know whether finding Substitute Issue B would be easy or hard. After all, ACA wasn't really a "National Socialist" scheme with "death panels." So no health care bill, and maybe those slurs or other similar ones get thrown at something else.

Basically, it seems to me that their finding supports two stories, both absolutely fascinating.

One is that health care reform, as enacted in the ACA, was a uniquely dangerous issue for House Democrats, and supporting it may have cost them the election -- through the mechanism they describe, in which supporting the bill gets translated into "too liberal."

The other is that under some conditions (such as the election of a new Democratic president and unified Democratic control), Republicans are able to use one issue to demonize mainstream Members of the House as "too" liberal. It's a limited ability; it works only when both the national information environment and local campaigns are making the same point. But the particular bill selected is incidental to the effect.

I don't see any way in what they've reported to distinguish between these two stories, both of which seem entirely plausible to me. Especially since one might suppose that the part about the GOP-influenced national information environment would be more effective in 1993 than 1977, and more effective in 2009 than 1993, given the rise of the GOP-aligned partisan press. I expect we'll get more research that with any luck will clarify some of this going forward.


UPDATE: Don't miss Eric's response below.