Anybody who's been reading Nate Cohn at the New Republic the past couple weeks could be forgiven for thinking he's moved on from polling analysis to devote himself to a new passion: concern trolling. The message his pieces have constantly been giving off, at least implicitly, is that "Democrats have no chance of retaking the House in 2014, so not only don't get your hopes up, but don't even try." He started off on September 26 arguing that Wendy Davis has no chance of winning the Texas governor's race (an assessment that, to be fair, I agree with), followed by a series of articles explaining why the government shutdown won't be that bad for the GOP/could be worse/they'll recover/whatever. Next came a couple of posts explaining why recent PPP polling (featured on this site) does not portend a Democratic House takeover, and arguing that Democrats will not retake the House even if they win the popular vote by 8 points. It's this latter article I chiefly want to respond to, but first let me say a few things:
1. I do intend to make the case for optimism regarding Democrats' 2014 prospects in the House, but I am not saying I think Democrats will retake the House, or even that they are likely to do so. I just think there is a greater chance it will happen than Nate Cohn and a lot of other commentators/prognosticators. Which brings me to a second point:
2. I don't really mean to pick on Nate Cohn (although in any case I doubt he cares what an anonymous, extremely sporadic DKE diarist thinks about him.) I just chose his articles because unlike Charlie Cook, Stu Rothenberg, etc, I follow him on twitter and see when he publishes new material. I'm sure he does not actually intend to come off as a concern troll.
3. It is kind of strange to see Nate's pessimism re: 2014 while John Judis is apparently predicting the death of the GOP on the same website.(1)
So with that out of the way, let's take a look at what Nate had to say most recently:
1. Democrats are "far from likely" to take back the House in 2014 even if they win the popular vote by an 8% margin. Instead they will "probably" need a double-digit margin to do so.
2. Democrats can't repeat a "2006-esque" wave because prior to 2006, the GOP held 18 seats with a PVI of D+2 or greater, and now they only hold 2. In other words, there are many fewer Democratic targets.
3. Polling does not show Democrats with the double digit lead they need to win back the House, and since the GOP can't sink any lower than they are right now, they won't get it.
Now, I do not intend to argue with his third point. I can't predict the future, neither can Nate Cohn, and neither can anyone else. On November 3, 2004, who would have predicted that 2 years later Democrats would retake the House and Senate, and 4 years later Barack Obama would be elected President? Absolutely nobody. I have no idea what polling numbers will look like a year from now, or what will happen in the US and the rest of the world to move those numbers. So maybe Nate's right on this point, and maybe he's not; I don't think it's really worthwhile to engage. Instead, what I do want to argue is that, if Democrats' polling numbers do hold up, and those numbers translate to a victory by 8% on Election Day 2014, or even by 7%, or even by 5-6%, Democrats will retake the House. Here is my thinking.
Let's start with a little bit of history. Since 1960, only twice has a party won a majority in the House while losing the popular vote: in 1996, when the GOP lost the popular vote by .07% but won 227 seats to the Democrats' 206, and last year, when the Democrats won the popular vote by 1.4% but took only 201 seats to the GOP's 234. For the Democrats to win the popular vote by 5% or more and still not win a majority of seats would be historically unprecedented. Of course, that doesn't mean it can't happen, but it should give all those pessimists out there a bit of a pause. Secondly, since 1960, 7 elections have seen one party gain 4.5% or more of the popular vote. In these 7 elections, the smallest number of seats gained was 26.(2) This number would, of course, be more than enough for the Democrats to take back the House next year.(3)
Now, let's take a look at what's out there in terms of pickup opportunities. Nate argues that in 2006, Dem pickups were dependent on seats with PVIs of D+2; of which the GOP held 18 prior to the 2006 elections but only 2 now. Therefore, "the best pick-up opportunities are already held by Democrats," and "a 2006-esque wave would only barely get the Democrats over the 17 seat threshold they need to take back the House in 2014." I think this is just plainly wrong.
Nate based his analysis on the PVI of the seats the Democrats took in 2006 and the PVI of Republican held seats today. If instead of PVI you look at the percentage that Obama took last year, we get some pretty different results. By my count(4), the GOP holds 5 seats where Obama took 52% or more of the vote. It holds 13 seats where Obama took at least 49% but less than 52% of the vote. And it holds 28 seats were Obama took at least 46% of the vote but less than 49%. Those numbers look quite different: if the Democrats can win half of the GOP seats where Obama took at least 49%, and 1/3 of the seats where he won between 46% and 49%, they'll have more than enough to retake the majority. It's not likely or easy, but it is certainly doable. If Democrats win the popular vote by 5% and don't do it, I'd be shocked; if they won by 8% and don't do it, I'll cover my apartment with framed Nate Cohn articles.
I think part of the issue here might be with PVI itself.(5) As far as I understand it, PVI is basically an indication of where a district stands in relation to the past 2 Presidential elections. So, in simplified terms, if the average GOP vote in the past 2 Presidential elections was 52%, and a particular district voted 51% for the GOP, that district would have a PVI of D+1, since it is 1 point more Democratic than the nation as a whole.
Where am I going with this? Well, in 2006, the GOP had won the past 2 Presidential elections(6), and the GOP vote in 2004 and 2000 averaged out to 49.3%. So in 2006 an R+1 district would, on average, have given 50.3% of the vote to GW, an R+2 district would have given him 51.3% of the vote, etc. However, in the past 2 Presidential elections, the GOP has averaged only 46.5% of the vote. So now even an R+3 district will only have voted, on average, 49.5% for Romney.
All this is just to say that not all R+3 (or R+2, R+4, etc.) districts are created equal. If the country is voting 60% Democratic on average, even an R+7 district is still going to be within reach for the Dems. I think, though I'm not positive, that what Nate has done here is compare districts' current PVIs not with a modern PVI (ie calculated according to the 2012 and 2008 elections) based on their former borders, but with their 2006 PVIs-which were calculated according to the 2004 and 2000 elections. And if so, that's a problem. Because today, a district with a PVI of R+2 is a district that Romney and Obama about tied in, and that, all else being equal, Dems should have a pretty good shot at. In 2006, a district with a PVI of R+2 was a district that Bush beat Kerry in by about 4 points. Basically, I just don't think you can compare PVIs at different points in time like that, because the meaning of PVI is constantly changing.
That (hopefully) coherent aside on PVIs concluded, what's particularly strange here is that back in April Nate Cohn agreed with me!(7) I'd love to know what's changed his thinking since then.
(1)Side note: I haven't read Judis's article. Is it any good?
(2)By the Democrats in 1982.
(3)FYI, admittedly, 1960 was a completely arbitrary choice.
(4)Based on the Pres by CD numbers available on this site.
(5)I don't claim to be an expert, so maybe I'm understanding PVI incorrectly.
(6)Well, they really only won 1, but I digress.
(7) "Democrats could probably win the House with a modest victory in the House popular vote, perhaps something like the 4 or 5-point margin suggested by a FiveThirtyEight model in 2011."