Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

How high is the House levee?

October 17th, 2013, 7:33pm by Sam Wang

Today we have some outlier statements made by other political analysts. As rounded up by Andrew Sullivan, we have Nate Cohn at the New Republic quoting Alan Abramowitz:

Democrats need a 13 point Democratic edge on September 1 to win the 17 seats necessary to retake the chamber in November.

For any wave election, that would be a very soundly built levee. However, recent data suggest this statement is rather overstated, for two reasons:

  1. In the last 6 elections, the drift from previous-year generic-opinion polls to November outcomes is a median of 3 points (4.6 points for midterms only).
  2. Republican-favoring gerrymandered districts are weaker than they look. Because they seem to have more independents than other Republican districts, the necessary national popular-vote margin for a Democratic takeover is lower than in 2012, and I estimate it at about 4 points.

I have analyzed this all recently. I think the levee could be breached by a 7-point generic polling margin next October. Also, here’s a rather good overview by David Wasserman at the  Cook Political Report. Let me review a few points about what appears to be in past and present polls.

Regarding point #1, Abramowitz’s claim runs up against evidence showing that midterm generic Congressional races are more stable than is commonly supposed:

An exception is the wave election of 2010, in which the race moved about 6 points toward Republicans in October:

To revive an old promise I made last year: if the race moves that far toward Republicans during the month of October 2014, I will eat a bug.

Does the generic opinion predict actual votes? To answer that, we have to quantify Delta, the 12-month difference between opinion and the actual national vote. Using generic Congressional ballot data from RealClearPolitics, I get the following medians:

Year 1 year before Outcome Delta
2002 D+0.0% R+4.6% Toward R 4.6%
2004 D+1.5% R+2.6% Toward R 4.1%
2006 D+8.0% D+7.9% Toward R 0.1%
2008 D+12.0% D+10.7% Toward R 1.3%
2010 D+1.0% R+9.4% Toward R 10.4%
2012 D+1.0% D+1.3% Toward D 0.3%
Average D+3.9% D+0.5% Toward R 3.4 +/- 4.0%

The correlation between polls now and eventual outcome is r=+0.89, which is pretty good! Indeed Democrats are likely to lose ground, but by less than implied by other analysts. The average movement is Delta=3.4% toward Republicans, with a standard deviation of 4.0%. (For total nerds: my discussion of the nice work of Bafumi et al. is here.) I note that the drift over the final two months of the campaign, from September to November, should be smaller.

In regard to point #2, here is a mathematical fact about redistricting: when you pack your opponents tightly into a few districts, the swingable voters are more likely to be in your districts. Call it Gerrymandering Lemma #1.

Here is some evidence for Lemma #1:

(An aside: The lemma is even mathematically provable! I estimate that as a general rule, win margins in a gerrymandered district should be discounted by as much as one-third, i.e. a 15-point win is only about as safe as a 10-point win elsewhere. I’ll write that up at some point.)

To summarize: Democrats face an uphill battle for House control, but the levee built by gerrymandering is not so high as you have been led to believe. A popular-vote win of 4-5 points might be enough. In terms of opinion polls, which usually overestimate Democratic turnout, if the generic ballot is at D+7% next October, it might be time for Republicans to start looking for high ground.

Update: Nate Cohn and I find points of agreement. See comments.

Tags: 2014 Election · House

36 Comments so far ↓

  • MAT

    So, if I’m following all this correctly – to simplify this into something I can explain to almost anyone: Dem’s should really focus on districts with a combination of lowest R margins but the largest percentage of Dems + Independents?

    Or am I oversimplifying?

    • Sam Wang

      Basically yes, but the problem is identifying independent-rich districts. My preference would be to formulate it like this: focus on (1) all R districts with margins less than 10%, and (2) all R districts in FL, MI, OH, NC, PA, VA, and WI with margins less than 15%.

      Note that if conditions are anything like they are now, I would give the exact same advice to Republicans. That’s the nature of knife-edge advice.

    • bks

      Is that the best we can do, Sam? It seems like we should be able to do realtime (for slow values of realtime) analysis of where the GOP is spending money to shore up districts and mobilize spending in districts that are ignored. Or perhaps we should spend a small amount in a district to force the GOP to spend there and then quickly shift our resources to a second district.

  • Amitabh Lath

    The “Independents moving to D” narrative seems to be gaining currency, which is more reasonable than hardcore right wing voters changing.

    But how does this affect the Republican primaries?

    If anything, it will move them even more right as the right-center moves away. That will either force the current republican congressman to the right or elect a far-right candidate to stand against the Democrat in November.

  • Nate Cohn

    I don’t think we really disagree. Elsewhere in the article, where I’m actually outlining my views, not Abramowitz’s, I argue that a 7 point advantage in the generic ballot would lead to a long election night where either party could takeover. That’s the same as your position, at least as I interpret it.

    • Sam Wang

      Maybe we’re not far apart. This is an interesting problem in parameter estimation. I found his estimate a little shocking, that’s all.

      The magic popular vote threshold seems like it should be in the 4-6% range. Could be that I am missing something – the “packing of independents” concept had escaped me until last week when the data made me think about it harder.

  • Nate Cohn

    I’ve disagreed quite a bit with some of your other work on the House. But on this particular question, I think we’re quite close, both absolutely and with respect to the others thinking about this stuff. I think, for instance, that the other person who thinks the popular vote threshold is lower than the uniform swing, Harry Enten, is more skeptical of the argument that the generic ballot is biased toward Democrats.

  • Amitabh Lath

    I am still not comfortable with the size and speed of these swings.

    The first derivative is so damn large, it beggars understanding. We don’t know who is driving these red districts blue (probably independents turning D, but confirmation?)

    Extrapolating to November 2014 is a fun project, but that’s all it is at this moment. To make it worse, most of the calculations do not discuss uncertainties (except Sam, who always gives the standard deviation).

    Seriously, all this is happening due to a few thousand people in a dozen or so districts changing their minds. Within a timescale of a week.

    Journalists should go out there and sit and talk in diners and daycares and coffeeshops and figure out what this is. Is this a spike that decays quickly as the football season gets going, or a slow realization among lifelong Republicans that they’ve allied themselves with nihilists and want out (the “I’ve made a huge mistake” explanation).

    • me

      “Is this a spike that decays quickly as the football season gets going, or a slow realization among lifelong Republicans that they’ve allied themselves with nihilists and want out (the “I’ve made a huge mistake” explanation)”

      The latter.

      They’ve probably poisoned the well one time too many.

      I’m sensing an undertow that we haven’t seen in quite some time: a fundamental frustration with the too conservative, dysfunctional Republican party that rivals the frustration with too liberal, dysfunctional Democratic party as the Carter admin ended.

      Good polling and analysis of the self-described independents should be able to assess the validity of this hypothesis.

    • Olav Grinde

      I wonder whether “independent” is too nebulous a term? What would be a reasonable definition — people who don’t blindly always vote one party? The voter’s party registration?

      Or is it more appropriate to say that, in this day and age, there is a very large percentage of voters that have zero party loyalty?

  • Olav Grinde

    Amitabh: “…lifelong Republicans that they’ve allied themselves with nihilists and want out (the “I’ve made a huge mistake” explanation).

    The Houston Chronicle recently expressed deep regret at having supported Ted Cruz. They acknowledged their mistake.

    On another forum (The DailyKos), I suggested the Chronicle do penance by supporting Wendy Davis for Governor.

  • Amitabh Lath

    Me, Olav: Deep in my heart I desperately want to believe that there is some wave of rationality is sweeping through deep red districts.

    But I think not.

    But the speed of the change makes me believe this is a blip, and by Thanksgiving we will back at status quo ante.

    Opinion in these red districts will snap back soon.
    Rather than poisoned well, the whole shutdown and default debacle will come to be seen as a quixotic and brave effort, just taken to far.

    As Olav points out, “independents” is a squishy term. Their positions on the Affordable Care Act, Immigration, taxes, global warming… are all still aligned with Republicans.

    They may have said one thing to pollsters in the heat of the moment, but anyone who changes their mind that quickly can/will change it back just as quick.

    Olav: as for the Houston Chronicle’s takeback of the Cruz endorsement, the NYTimes has an article today about how Cruz still has deep support among Texans. If the Houston Chronicle starts to see subscriptions decline, watch them take back the takeback.

    • mediaglyphic

      At the very least the polls show us that many seats are vulnerable.

      I wonder if an effort to register voters in these districts combined with better candidates (a point Nate Cohn keeps making) and better technology (Narwhal), can take the house back.

    • Amitabh Lath

      Hey media, I don’t mean to be a gloomy gus.

      It’s just that sharp edges on plots are almost always due to the measurement device rather than the system being measured.

      Some of us in central jersey should move to Cherry Hill and run in NJ2 and get that back to D.

  • mediaglyphic

    If it keeps on rainin’, levee’s goin’ to break,
    When The Levee Breaks I’ll have no place to stay.

    Mean old levee taught me to weep and moan, [X2]
    Got what it takes to make a mountain man leave his home,
    Oh, well, oh, well, oh, well.

    Don’t it make you feel bad
    When you’re tryin’ to find your way home,
    You don’t know which way to go?
    If you’re goin’ down South
    They go no work to do,
    If you don’t know about Chicago.

    Cryin’ won’t help you, prayin’ won’t do you no good,
    Now, cryin’ won’t help you, prayin’ won’t do you no good,
    When the levee breaks, mama, you got to move.

    All last night sat on the levee and moaned, [X2]
    Thinkin’ about me baby and my happy home.
    Going, going to Chicago… Going to Chicago… Sorry but I can’t take you…
    Going down… going down now… going down….

  • Trudy Peck

    There are two good reasons why Democrats loathe Scott Rasmussen: he has been very accurate and very early in predicting the electoral whipping that Democrats have incurred since the New Jersey and Virginia gubernatorial races in November 2009. His most recent poll of likely voters brings more bad news for Democrats: by 55 to 35% Americans support the repeal of Obamacare. Since June 2011, Rasmussen has taken 21 polls on Obamacare’s repeal. In each poll, repeal has won by double-digits.

    • Sam Wang

      The claim that Rasmussen is some kind of standard of accuracy is so risible that I simply had to let it through. Here is a recent survey showing sentiments on the Affordable Care Act, also known as Obamacare. Note that about one-fourth of the opposition wants something more liberal.

    • Craigo

      1. Scott Rasmussen no longer works Rasmussen reports.
      2. Rasmussen Reports employs questionable methodology which results in mediocre accuracy.
      3. No one pollster should ever be trusted against the field. Until more houses show a double-digit lead, we should assume that one probably doesn’t exist.

    • Craigo

      I somehow got 3 mixed up with another comment about the VA-Gov race on a different blog. Apologies.

      I’m also amused by the idea that the 2012 elections either did not occur, or constituted a whipping for the Democrats. If the latter is true, God only knows what’s going to happen to the House GOP when they lose the national vote by more than a point or two.

    • Todd Horowitz

      @Craigo: Republicans certainly act like the 2012 elections did not occur.

    • Ross C

      So awesome. The latest Rasmussen poll for the Virginia governor’s race in a couple of weeks has McAulliffe 50, Cuccinelli 33. Oh the wailing of Democrats! Curse you Rasmussen!

    • Charles

      First, the latest Ras poll only has a -2% margin between favorable and not favorable. In your defense, it is polled often and therefore we could be looking at two different dates. However, when looking at the questions posed, they do not ask the critical question regarding why people don’t like it. When that question is asked in other polls somewhere between 5%-15% respond that it doesn’t go far enough. Therefore I do not believe it is the resounding issue that you do.

  • Dave M

    “Since June 2011, Rasmussen has taken 21 polls on Obamacare’s repeal. In each poll, repeal has won by double-digits.”

    That’s odd. Since June 2011, I too have taken 21 polls on Obamacare, and repeal has garnered less than 10% in every one of ’em.

    • Olav Grinde

      What Rasmussen doesn’t admit, are their secret, unpublished polls that show that three times as many Americans want to repeal the Republican Party, as repeal Obamacare.

  • Jinchi

    Replying to Olav Grinde above: I don’t think that the term “Independent” is nebulous at all. There are people who are partisans and there are others who are not.

    I think the confusion comes when political discussions conflate the words “liberal” with “Democrat” and “conservative” with “Republican”. Being independent doesn’t mean being void of opinion or ideology. And it certainly doesn’t mean that a person switches from party to party at random. It typically means that you value the policy over the party.

    • Craigo

      The vast majority of self-described independents are partisan in all but name. I.e., they’re not registered as Republicans or Democrats, but their voting behavior does not differ significantly from them. Independents are people who value party but don’t want to admit it.

      The proportion of true independents is estimated at 10% in presidential years, less in midterms. They’re typically unengaged, uninformed, and unlikely voters. It might be more accurate to say that they value policy just as much as they do party, but only because they care very little about either.

  • Amitabh Lath

    Sam, set the filter back to Defcon 1. The desperate tone of these posts hawking the latest poll from Rasmussen, or “analysis” from Heritage is hard to take.

    They are like a cargo cult with their bamboo airplanes and torch-lit runways hoping if they follow all the forms correctly, the big metal birds with goodies will start to land again. (Look! We got poll numbers! Cross-tabs! Skewness! Kertosis! Bayesian! Surely Romney is going to win now!)

    Anyway, what’s their fascination with Obamacare popularity polls? You need to determine how important this issue is to people. What fraction will base their vote on it?

    Don’t most people have health insurance through their employer? I suspect the ACA is of academic interest to them, at best.

    • Matt McIrvin

      Republicans believe that they’re going to lose their jobs or their existing coverage (employer or Medicare) because of the ACA, Republican employers have been blaming everything they do to their workers on Obamacare, and Fox News has been pushing the idea full-time. The reality of the law is of little consequence here.

  • Jinchi

    “The vast majority of self-described independents are partisan in all but name. I.e., they’re not registered as Republicans or Democrats, but their voting behavior does not differ significantly from them.”-Craigo

    Sorry Craigo, but this is the same nonsense argument I’ve been hearing from partisans for years now. It’s little surprise that a liberal will typically vote for the Democrats and a conservative will typically vote for Republicans. Those, after all, are the options available to them. It’s not because they are inherently partisan. As to your comment about “true” independents, I can only assume you’re refering to people who mark the ballot at random.

    A partisan Democrat will check off any D on the ballot and a partisan Republican will back any R. But a liberal might decide that there’s really no point in showing up at the polls to back a blue dog who wants to bomb Iran and a conservative might decide that she’s not willing to check the box next to some nut who goes on about “legitimate rape”. Likewise, Democratic partisans will often cheer the primary victory of a Tea Party nut, because it makes the final vote marginally more likely to swing towards their candidate. This of course implicitly assumes that there are actual independents, who are paying attention.

    We’ve been through several wave elections in the past few years because people are engaged and informed and don’t particularly like what they’re seeing from their elected Representatives. Those people are your independents.

  • Pat B

    Glad to see you are back at it. I enjoy your posts and predictions.

  • Jinchi

    “How exactly would you define an independent.”

    I’d define an independent simply as someone who doesn’t identify with either party. It doesn’t say more about a person’s views than that. It won’t tell you how often they switch parties or even whether they choose to vote. In fact, self-described independents can be registered Democrats or Republicans, since in many jurisdictions, you have to register on a party line in order to vote in a primary.

    “they’re not registered as Republicans or Democrats, but their voting behavior does not differ significantly from them.”-Craigo

    I assume we can agree that the voting behavior of Democrats differs significantly from the voting behavior of Republicans. The studies Craigo cites simply illustrate that independents span the ideological spectrum. This counters the common misconception that independents are centrists, or moderates, or people without well thought out opinions.

    We do know that a significant number of voters change their partisan choice from one election to the next. That’s why we’ve had several wave elections in the past decade. The distinction between partisans and independents comes into focus in the reason they would consider changing their votes. Partisan Republicans cheered the shutdown because they hoped it would damage the Democratic president, while a conservative independent might hate it because they want the government continue to function. This is certainly what the polling at the time seemed to indicate.

    Likewise, partisan Democrats celebrate the primary victories of Tea Party candidates because it makes it more likely that the Democrats will retain the Senate, while left-leaning independents cringe, seeing that the Tea-Partiers who did get through, now control a major branch of the government.

Leave a Comment