Catch of the Day: What Ebola Effect?
I'll award a Catch to Harry Enten, who knocked down the idea that the Ebola scare would deliver the midterm election to Republicans.
The theory Enten responded to involves trust in government: People are worried about Ebola, they don't trust the government to deal with it, and therefore they'll vote against incumbent Democrats. But "trust in government" is tricky. It mostly follows partisanship, as Enten demonstrates. The causal direction is that people are Republicans, see that Barack Obama is president and therefore don't trust government -- not the other way around.
This isn't to say that Ebola couldn't hurt the Democrats. Research has shown that, in some situations, bad news can hurt incumbents at the ballot box even when they have nothing to do with it (with examples being bad weather and losses by local sports teams).
In this case, however, nothing seems to be going on.
The best indicators for any anti-administration or anti-Democrat movement would be the HuffPollster averages for Obama's approval and for the generic House ballot (that is, a ballot with no names). The Ebola story hit the news in early August, then became a frenzy at the beginning of October. Obama's approval fell from 43.0 percent on Aug. 1 to 42.4 on Oct. 1 to 42.2 today. And if you switch to the HuffPollster setting that is more sensitive to short-term changes, the difference from August to today is just half of one percentage point. That's about as flat as it can be.
On the generic ballot for House candidates, Republicans moved from tied at the beginning of August to about a three percentage point lead today. Yet movement toward Republicans had begun much earlier in the summer. It's also tricky to interpret for technical reasons. And the trend is with midterm-election models, which anticipate voters finding some excuse for turning against the party holding the White House.
So while it is possible Ebola could be costly to Democrats, that doesn't appear to be happening by the numbers. And it almost certainly isn't happening as a result of declining trust in government.
Therefore: Nice catch!
I'm looking at the "moderate" setting in their custom charts. There's something screwy about HuffPollster's generic ballot charts; the custom one doesn't match the "HuffPost model" even with the settings all the same. The story I'm getting from the custom chart is consistent with the RCP polling average, so it's probably correct.
Pollsters switch from using registered voters to likely voters over this time frame, and the latter always is better for Republicans; that alone could account for most or all of the change from June to now.
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Jonathan Bernstein at firstname.lastname@example.org
To contact the editor on this story:
Katy Roberts at email@example.com