In the most recent post over at 538, Nate Silver makes a brief parenthetical statement about Pennsylvania that got me wondering if there was something bigger going on.
(So pronounced are the regional variations in Pennsylvania that Republicans were able to devise a redistricting plan in which Democrats won just 5 of the 18 seats for the House of Representatives this year despite their House candidates winning more votes over all throughout the state.)
And this was also true of Ohio, where Republicans won 12 out of 16 House seats despite losing the state’s popular vote to Obama. Those are huge lopsided Congressional advantage for Republicans in a swing state, especially ones with a Democratic lean in the popular vote. And the Republicans retained the House despite losing at most other levels across the country. This leads to the question: How pronounced is the difference between each states’ Congressional representation and its votes for President? (And if the discrepancy is large, what are the reasons?)
Maybe this has been covered elsewhere, but out of curiosity I wanted to see what the picture looked like for myself. I gathered two sets of data:
1) The state-by-state current vote totals for Obama vs. Romney (from the Cook Report spreadsheet).
2) Each State’s Congressional delegation percentage, from the 113th Congress state listing on Wikipedia.
Here’s the final CSV file.
I tried two ways to visualize the data to see what the current state of the discrepancy between the Obama/Romney state voting and the Congressional delegation percentages (i.e 5 Dems and 5 Reps would be 50%).
Way #1: By Popular vote margin
A couple things stand out here:
1) Only 1 state, AZ, had Romney winning the popular vote but Democrats holding a majority of the Congressional Delegation.
2) Several states had the reverse, where Obama won the popular vote but Republicans had a majority of the delegation.
3) Almost every swing state, including OH, FL, PA, VA, NV, CO, and NC have lopsided Republican delegations (Updated: not NH).
Way #2: By Quadrant
It does seem like the entire chart, which you would expect to center at 50% population vote matching with 50% congressional representation, is shifted several steps to the right. The discrepancy between B and C does not tell the whole story, but it does seem clear that the popular vote totals have to be very far in the favor of Democrats before the House delegations start to match (except one time, in Arizona).
—————-
There are several reasons this could be the case, here are the ones I am thinking:
1) Better Republican candidates at the House level across the board.
This seem unlikely because of the sheer numbers in every state. Having all those candidates randomly be better than their opponents at once is pretty unlikely.
2) Obama more popular than other Dems in general
I don’t like this answer because Dems did so well across the country in other state-wide races, including ballot questions and Senate races.
3) The Incumbency advantage
This one seems very plausible to me. Incumbents have huge advantage (fundraising, media appearances, recognition, etc) and there was a very big Republican wave in 2010 that left a lot of Republicans as incumbents for 2012.
4) Aggressive Gerrymandering of districts
Republicans across the country, swept into power in 2010, had the unique advantage in many states to use the once-a-decade census numbers to draw up new lines and redistrict states in a partisan manner. We in MA have a nice long history of this sort of shenanigan, I think at least one speaker has gone to jail for it?
If #4 was the case, then the Republican House majority could be around for a very, very long time.
Any thoughts on how could we determine if that’s the situation. or if its just a combination of the other 3? (or is something else going on?)
kbusch says
Both of NH House seats have flipped Democratic whereas, according to your chart, they’re both Republican.
sleeples says
Oops. I was updating everything from 2008 data, overlooked that 2-0 was not the same as 0-2!
Updated post
kbusch says
Whenever I discover I’ve made a mistake like that, I usually discover more mistakes of a similar nature. So you might want to check other states too.
What you’re doing here is very useful, and it’ll be more useful if it’s completely solid.
Christopher says
GOP did after all control most states following the 2010 election so that is not surprising. I believe I also heard that the total number of votes for all Dem House candidates nationwide exceeded that of the GOP.
sleeples says
Any thoughts on how we could quantify that advantage?
Seems like some way to calculate it could be useful in the general rhetorical war on gerrymandering and redistricting to be able to say how underrepresented certain areas are. But I’m not quite sure how a model to test that would work, or what data we would need.
jconway says
The only way to stop abuses is to get California style non partisan commissions. I would argue that this not only stops gerrymandering from being a bipartisan incumbent protection racket but also ensures more proportional representation overall. In California several moderate Republican challengers gave entrenched liberal incumbents an actual election and gave their districts a choice. While I am happy liberals liked Lois Capps won their races, I am glad moderate sensible Republicans like Abel Maldonado were able to make a race out of it. The voters won in that election.
In most states where Republicans were in charge of redistricting like PA, VA, MI, and CO the states are a lot bluer than they should be. Liberal enclaves like Fort Collins, CO and Charlottesville, VA should not be represented by hardcore conservative Republicans. But those liberal, urban, high population centers are split up across three conservative districts.
The Illinois redistricting round was heavily criticized by Republicans, especially those that lost. But it actually corrected prior abuses when Republicans were in charge of the process. It makes more sense that Bill Foster’s district united the urban Aurora and suburban Naperville into one district, where they had been split across several rural districts before. Its nitty gritty work and each district is unique.
stomv says
There’s no question that it’s the result of gerrymandering.
A number of those states have something in common:
* dense urban cores, often overwhelmingly black relative to the state as a whole (MI, OH, VA, PA come to mind; I don’t know about FL, WI, CO).
* lots of suburbs, often full of n-th generation state residents (MI, OH, VA, PA, WI; don’t know about FL, CO).
So you end up with the situation where it’s *easy* to pack Democratic voters in urban areas, and ensure that the suburban districts contain just enough GOPers to stay GOP. The effect is most pronounced in the ‘x2 election non-POTUS, but the ‘x2 election with POTUS still sees it. Note that PA, OH, MI, WI, VA, and FL all have majority GOP house, senate, and GOP gov.
You can use non-partisan district drawing, but you know another way: turn out the Democrats on the non-POTUS elections. Then you won’t have GOP controlled legislatures in Obama states so often, and then you won’t have the gerrymandering.
Until the Democrats figure out how to turn their voters out on the non-POTUS biannual elections (or, even trickier, in places like VA it’s on odd years!), this phenomenon will continue, and be most obvious on the first election after the census when the data is most fresh, thereby making the gerrymandering most effective.
P.S. Because 2020 is a POTUS election, it won’t be as much of a problem in the 2022 election, because the state legislatures will have “benefited” from a higher Dem turnout due to the POTUS election.
P.S. I’ve been griping about this Rust Belt phenomenon for weeks now. I’ll turn it around though: given that the GOP has the gov, state house, state sen, etc., why do they underperform for POTUS and, to a lesser extent, SOTUS?
jconway says
Until all states have the Calfornia model it behooves us to make sure that Delay style tactics stop once and for all. Again, Democrats are guilty of redistricting ploys, but usually, as my IL example shows, it also empowers urban and suburban districts that were split up into rural ones. GOP redistricting does the opposite and basically ghettoizes Democratic voters, especially in the south. 2020 will be crucial, especially if we want to push GA, SC, and AZ into the mix.
Between now and then getting state legislative races going is another key factor. I know in Chicago we have sent volunteers to canvass WI during the legislative recalls and we can do that again, VA will also be key.
mski011 says
If you ask NV politics guru Jon Ralston he might say that the only reason the Dems do not control the delegation 3-1 is due to crummy candidates. Either way, that state has a court map and it is probably best described as a 2-1-1. That is Dem, GOP, swing. That’s what we got. The delegation is split now.
I think Nevada is a rare case where the Dems are overwhelming geographically concentrated in Clark County (Last Vegas). Perhaps with a better economy again Washoe County (Reno) could push its Congressional district to swing, but the huge GOP rural areas make that impossible for now.
I think the Nevada example also explains how states like Pennsylvania get gerrymandered. The Democratic population is so concentrated that unless fair districts are drawn it is easy to gerrymander the ridicuous map they now have.
fenway49 says
is that, like Democrats, people in general are overwhelmingly located in Clark County. It has nearly 70% of the state’s population. Washoe, containing Reno, has 17%. That leaves less than 15% for all the other, very rural counties.
With four districts in the state and a reasonably strong Democratic brand in Clark County, the lines could be drawn in a way that favors Democrats more, but a court map’s not going to do it that way.
PA, as I recall, had this problem even in the last redistricting (2000). It went to the Supreme Court, where they declined to decide it as a “political question.” It seems to me this is the exact kind of thing you would NOT want the political branches to be deciding.
I was a law student when the case came up and the lawyers came to our school for a mock argument in preparation for the real thing. Bush v. Gore was fresh in everyone’s mind, and my thought was that, in the absence of an electoral college winner, each state’s House delegation would cast one vote for the next President.
That means each state’s House delegation, in itself, has Constitutional significance beyond just being a few individual seats in a much larger body. With gerrymandering, the possibility very much exists in a close Presidential election that a Democrat could win PA by nearly ten points, but the delegation would vote for the Republican. That also could happen in a one-district state, but generally the Pres. vote and House vote goes the same way.
In any event, GOP strength in the Pa. House delegation was disproportionate from 2000 to 2010, and it’s much worse now thanks to 2010 Tea Party debacle.
bluewatch says
Look at it this way. We already know how republican state legislatures tried to suppress the vote with voter ID laws, reduced time for early voting, reduced time for registration, and new challenge laws. So, it should be no surprise to see that these same republican legislatures also Gerrymandered their congressional districts to their advantage. In particular, there was serious Gerrymandering in Florida, Texas, Pennsylvania, Ohio, and Arizona.
On this topic, Karl Rove is the actual winner. He funded many republican legislators throughout the country in 2010, while Dems were sleeping.
It’s not too early to realize what happened and develop a strategy to respond.
stomv says
The Dems went all in on health care. They ignored taxes, climate change, women’s issues, labor, education, and lots of other things to go big on health care.
The Dems won health care.
It came at a cost though — a major backlash. Tea parties and whatnot. As a result, we had a really tough mid-terms. We got whacked in the US House and the state legislatures. We weren’t sleeping — we got taken to the woodshed, in exchange for the Affordable Care Act. It was a trade.
Was it worth it? I’m not sure.
bluewatch says
We did not turn out our vote in 2010. We did not sell the benefits of Obamacare or the stimulus. We did not make efforts to elect Democratic state legislatures in many states. Massachusetts was an exception, due to the leadership of Deval Patrick and John Walsh.
stomv says
We didn’t turn out the vote? I suggest you compare turnout to 2006 and 2002 if you want to back up your claim with facts. I suspect you’ll find turnout to be pretty flat.
Sell the benefits of Obamacare of stimulus? We had all those damned “town hall” meetings to do just that. It wasn’t for a lack of effort, that’s for sure. We may not have performed well there, but we weren’t “sleeping.”
Make efforts to elect Democratic state legislatures in many states? What evidence could you possibly have of that? I’m all ears.
fenway49 says
is that the Voting Rights Act of 1965 (VRA) requires the creation of majority-minority districts in Southen states to protect minority group political power. In the Congress currently ending, the GOP held 102 House seats in thirteen Southern states, the Democrats 40. 24 of the 40 were VRA districts.
(BTW, so conservative is the native white Southern population that 9 of the 16 remaining Democrats were Blue Dogs. The other 7 represent Austin, Texas and places with huge Northern transplant population (DC suburbs in Virginia, Palm Beach/Fort Lauderdale, NC Research Triangle).)
Some advocates have argued that, since Democrat in today’s South largely equals African-American, the GOP is using the VRA to justify “packing.” For those unfamiliar, “packing” is when as many of one group are jammed into a district as possible. That makes that district very safe for that group (90% +), but means they have little chance of winning in any of the other districts.
These activists have said African-Americans might have more political power with a still-safe district of 65%, plus a neighboring district of 40% or so that they could compete in. This maintains the likelihood of at least one seat won, but increases the chances for two. The ever-helpful Roberts Supreme Court, however, held in 2009 that, where members of a minority group do not constitute a numerical majority, the VRA does not require allowing them a fighting chance to win political power by joining with like-minded white voters. Barring state law to the contrary, legislatures are free to do it, but they don’t have to. Most GOP legislatures won’t, obviously.
(The other strategy, “cracking” is when the redistricters reduce a political group’s voting power by scattering them across as many districts as possible. If a state’s party breakdown in 60-40, the majority can make each individual district 60-40, making it hard for the minority party to win in any of them. To some extent this has happened in Massachusetts.)