Sean Trende has a nice item exploring the worst-case-scenario for Republicans in this year's Senate races. I agree with him that most forecasters underrate the chances of an extreme outcome in either direction. What about a possible continued recovery of President Barack Obama's standing?
According to the HuffPollster estimate, Obama’s approval rating bottomed out at 42.5 percent in the first week of December. He’s up to 44.4 percent now, a 2 percentage-point gain over about four months. We’re seven months from Election Day; if the trend continues, Obama could get up to about 48 percent. Of course, the trend could just as well reverse or accelerate.
I’m somewhat skeptical that it would make much difference. Presidential approval can have direct effects on voters. But it also operates indirectly, through candidate recruitment, donor generosity and other forms of party enthusiasm. Candidate recruitment is a huge part of that, and many of those deadlines have already passed.
Four presidents have made it to a second midterm in the polling era. In January 1958, Dwight Eisenhower was at 60 percent approval and was at 52 percent in early November. Ronald Reagan held steady, going from 64 percent in early January 1986 to 63 percent in late October. Bill Clinton was at 59 percent in early January and ended October at 66. And George W. Bush began 2006 with 43 percent approval and slumped to 38 percent by early November.
One reason to be cautious is that while we have estimates of the effects of approval, we only have four previous cases of an incumbent president and a second midterm. We have quite a few more midterms overall, and there may be nothing special about an incumbent’s second midterm. There’s just no way to know if there are particular effects that might play out in this situation (for example: does direction of the trend matter, or just the level?), though even if there are, they’re likely to be fairly small.
If I had to guess, I’d predict that Obama will continue along his current slow recovery path, barring disruptive events. That still makes him a drag on Democratic candidates, especially in Republican states, though less of one than Bush was in 2006.
Technical note. Unlike the simple average of the most recent polls at RealClearPolitics, Pollster estimates a trend line, using all the polls they collect. For that reason, new polls not only shift the current estimate, but the entire recent trend line. Therefore, it’s still possible that the early-December numbers, or any others, could change.
Don’t get too attached to those exact numbers; especially in the older cases, Gallup wasn’t polling that often, and random noise could be responsible for 5-point swings in either direction. It wouldn’t surprise me if Reagan, in particular, had actually slumped by Election Day, which would make his off-the-cliff numbers after Iran-Contra broke a little less dramatic, and would also sort better with his midterm losses.
To contact the author on this story:
Jonathan Bernstein at email@example.com
To contact the editor on this story:
Max Berley at firstname.lastname@example.org