Help with district-partitioning calculation?
(Considering that this is a fairly narrow-appeal post, I will pipe it over to the right-hand “Meta-Analysis” column shortly.) Dear PEC rea...
Gerrymandering creates a point of weakness
In the fight over the shutdown, a residual feature of partisan gerrymandering has become unexpectedly important. Andrew Sullivan quotes Kyle Kondik at Larry Sabato’s Crystal Ball:
[I]t’s the House Republicans in marginal districts who could see their ranks decimated, just like the House Democratic moderates whose anti-Obamacare votes couldn’t save them in 2010.
An examination of the MoveOn/PPP data suggests that this burden might fall harder on some Republicans than others. Perhaps surprisingly, a major cause seems to be partisan gerrymandering.
The white symbols show districts in states that I demonstrated in February to be gerrymandered to allow Republicans to eke out many small victories: FL, MI, OH, PA, VA, and WI. The swing in those 12 districts is 14.1+/-8.5 % (mean+/-SD), as opposed to 7.8+/-4.8% for the other 12 districts. (My analysis spreadsheet is here.)
It is a common fallacy to believe that seats gained by partisan gerrymandering are safe seats. In fact, the converse is the case. Gerrymandering achieves a net gain of seats by packing the opposition party into as few districts as possible. North Carolina is a particularly obvious example:
Representatives who benefited from the great partisan gerrymander of 2010 were given enough of an advantage to get into office narrowly. In a district designed to give Republicans a narrow advantage, Republican loyalists are likely to be spread thinly, with the balance of the needed votes being drawn from independents. Some of these independents might be more prone to anger about the current situation. These polls suggest that Republicans in those states might be particularly ripe targets for pressure.
The 12 districts plotted in white are FL-02, FL-10, FL-13, MI-01, MI-07, MI-11, OH-06, OH-14, PA-07, PA-08, VA-02, and WI-07.
Topics:
So the obvious answer for Democrats is to get more people in the gerrymandered districts registered and voting.
The republicans have thought ahead on that one with all the voter suppression legislation they’ve been pushing through.
An interesting analysis of the situation would be to look at the perfect storm of gerrymandering, media consolidation, suppressing lawmaking and those in combinations with other tricks up the GOP sleeve to try to stay in power.
The other interesting side effect is that you’ve tended to give your opponent a smaller number of very safe districts, which they can use as a base, while you’ve spread yourself thin to achieve your majority in many of the rest. I wonder if that has any impact on the D’s current (apparent) ability to stay united.
My impression was that this tendency contributed to the wave election in 2006, when even small movements in these finely-balanced districts could have an outsized effect.
The post article was nice but this analysis needs more press. If the republicans really begin to realize they could be at risk, then the behaviour might change.
It might be useful to distinguish between the two gerrymandering strategies: ‘stacking’ your opponent’s votes high in safe seats and ‘cracking’ your opponent’s vote into narrow minorities across seats. It’s by no means necessary to use both at the same time, and they behave quite differently in the face of changing voter preferences. ‘Stacking’ is quite resillient against unfavourable swings but has (comparatively) low potential maximum efficiency in realistic situations. ‘Cracking’ can win you a majority of the seats on 25% + 1 of the vote, but it’s much more vulnerable.
Sam, I am looking at your spreadsheet (http://election.princeton.edu/wp-content/uploads/2013/10/PEC_moveon_ppp_9oct2013.xls) and am wondering about the swing calculation (Columns F and G).
You are subtracting the 2012 R election margins from the moveon poll numbers. But I don’t think that poll reflects the actual midterm turnout D/R/I ratios (older, more conservative, etc).
If we assume the entire swing is due to independent voters (ie, D’s and R’s never change their minds) then we can go back district by district, and apply the correction to I voters and see how the result changes.
Then districts with large I populations will be more likely to change, which does make intuitive sense. Or is this not a significant effect?
Sam, any chance you could take Pennsylvania as an example? It would be nice to see an analysis of the type of gerrymandering in this state, and how vulnerable you deem the various seats to be in the 2014 election.
Certainly Pennsylvania is a fine example of gerrymandering: http://elections.nytimes.com/2012/results/states/pennsylvania.
Four out of 13 Republican representatives won with 55% or less of the vote, and another 5 with 55-60%.
Having been gerrymandered right out of PA-07–once a Republican lock, and more recently a seat which has seen a couple back-and-forth swings–and into PA-01, where the Power of My Vote is almost nil, I’m an example of the district packing Sam refers to.
In PA in 2012, not only were the five Democratic victories staggeringly lopsided (note that four of those five finished with a higher vote percentage than ANY of the 13 Republican victors), but the state as a whole contributed to the Democrats’ national popular vote victory in House races: PA Democratic House candidates received 50.28% of the vote to the Republicans’ 48.77%, yet won only five of 18 seats.
The North Carolina gerrymander is robust. Barring local scandals, only two Republican districts in NC are at risk in 2014. The Democratic coalition relies on low propensity voters. This is the main reason democrats got crushed in 2010. Democrats gained seats in 2012 despite a less favorable map because of turnout. They will be lucky to gain any seats in 2014.
David Abbott, which two districts? I live in Asheville, where our huge D majority has been left with no representation after the gerrymandering split our county between 10 and 11. As someone who has worked locally for the D party in ’06, ’10 and ’12, my gut tells me to at least hope that an angry and motivated electorate has a chance in to take 10 and/or 11 in 2014. Any thoughts?
Anger is an important component of voting. In 2010, the Republicans were angriest. In 2012, I think the Democrats will be the ones who are most upset.
One thing that is often forgotten in this discussion is that the districts are based on the 2010 census. The farther we get from 2010, the more demographic change has occurred. I’d be interested in an analysis of demographic change in swing districts post 2010.
Unfortunately, the example provided doesn’t support the thesis. While the 3 of 4 Democratic seats were, indeed, won by a landslide, no Democrat came within 6 points of the Republican in any currently R-held district (which are generally considered solid margins of victory). 7 of 9 R-held seats were won by more than 10 percentage points. Only one victory was actually “narrow” — a Democratic seat won by .2%. Without Obama coattails to ride, this seat will likely flip to R next year.
I’m sure that there are other states that could be used as examples that would better illustrate the point.