Thanks to their capturing the the state legislature in the 2010 wave elections, North Carolina Republicans were able to gerrymander the legislative districts, meaning they were able to draw the district lines in a way that gave them a considerable advantage going into the 2012 elections. While I have previously analyzed these districts in detail and just how strong of an advantage they confer upon Republicans and what that means for the 2014 elections, I now wanted to look at how the 2012 elections might have played out under a non-partisan or fair map and what that might portend for 2014 as I previously did for the US House of Representatives. That diary found that gerrymandering likely cost the Democratic Party control of the US House despite a modest popular vote win. There is good reason to believe that, once seats uncontested by one of the two major parties were taken into account, that the Democratic Party also won the popular vote for the North Carolina Senate and House though fairly narrowly. As such, this diary will look at how a non-gerrymandered map might have impacted the outcome.
In designing a map, there were several factors to consider. First and foremost was compliance with the federal Voting Rights Act which, while recently constrained by the recent Shelby v. Holder Supreme Court decision, still has a considerable power over the drawing of electoral districts in the United States. Thus this map would have to comply with the VRA, specifically concerning majority minority districts. Secondly, the North Carolina state constitution and US and State Supreme Court decisions have placed certain limitations upon redistricting, such as prohibiting excessive county splits and that no district may deviate from the mean by more than +/- 5%. In addition to that, it seems logical to limit excessive splitting of local jurisdictions such as cities and towns. With those constraints in mind, the districts I drew were to largely conform to existing Communities of Interest based on cultural, geographic, and demographic factors. Lastly, I fortunately had a good guide for what was a legal, largely non-partisan starting point with the previous decade's senate districts, while there were also multiple proposals for new non-partisan maps offered from legislators in the General Assembly during the 2011 debate over redistricting.
So let's take a look at both the actual map and the one I have drawn using Dave's Redistricting App, the data behind it, and what we can derive about the impact that gerrymandering had in 2012 and what role it might play in 2014.
These are the 50 state senate districts the Republican controlled legislature passed when they gerrymandered both chambers. You can find more detailed maps here.
Here I've colored the districts by the difference between the average vote percentage won by the statewide ticket in each district compared to the statewide average itself. You can see that Republicans have a fairly strong advantage and the vast majority of he state appears in varying shades of red. 33 districts were more Republican than the state in 2012 while 17 were more Democratic; Romney carried 33 to Obama's 17. Unsurprisingly Republicans won 33 seats to Democrats' 17 (though two were mismatched). This is despite a 1% popular vote win for the Democratic Party when uncontested seats are accounted for and means Republicans won roughly a 17% higher share of the seats than their share of the vote.
Graphically you can see just how one sided the chamber is and how Republicans were able to manipulate the process to build in an advantage for themselves. The 26th least Democratic seat, which theoretically would have flipped control of the chamber, is 5.7% more Republican than the state as indicated by the black line on the graph. Very few districts were even all that close to the statewide average. Note that I prefer to use the average of the 9 downballot statewide races when determining a district's competitiveness or political lean because there is a higher degree of statistical correlation between those results and the state senate election results than there was with the presidential results in reality (R-squared of .95 vs .89).
click through for full size
In this attempt at a non-partisan map I've undone the most egregious Republican gerrymandering such as with the districts in the major metropolitan areas, but also in a many of the rural districts where counties were strategically placed to pack or crack areas of Democratic strength. Let's take a look at the numbers behind it.
Moving left to right, you'll notice that all districts are within the 5% deviation from the mean, and 9 districts are majority non-white among the adult population while an additional one is significantly non-white enough to elect a minority for VRA purposes. Politically I've given the Obama/Romney vote shares of the two party vote and how each district compares to the national vote share percentage. Obama won 18 districts and did better than his statewide percentage in 19 total. The bigger change is in the downballot average where instead of 17, a total of 22 are now more Democratic on average than the state while an additional district voted for the Democratic ticket on average.
When we map these two things out, the state looks quite different:
What should become apparent when you look at these non-partisan maps and the previously shown version of the actual districts is that there are several more districts that are closely divided between the parties. Graphically the distribution, while nonetheless favoring Republicans slightly, is still much more evenly balanced than the actual map.
Here the 26th least Democratic district is just 2% more Republican than the state on average, resulting in a nearly 4% shift away from the Republican Party. Importantly, there are three times as many districts that are within 2% of the statewide average and nearly twice as many within 4%, resulting in a much larger degree of competitiveness in the chamber. While the actual state senate elections had only 2 races out of 50 where the outcome was decided by less than 5%, under this map there undoubtedly would have been several more close races.
Since there are only 50 districts in the chamber and many of these either weren't changed at all, or are too one-sided in favoring one party over the other, I'll go down the list of key districts to assess how this map might have changed the 2012 result and how the races had been run.
SD-01
The 1st district in the northeastern Outer Banks changes only slightly, but given how this race was by far the closest of all state or federal elections in North Carolina at a margin of just 21 votes, the change is both crucial and leads me to believe the case for this district flipping back to the Democrats is ironclad. Here I've reverted the district back to a version similar to the 2010 district by restoring Tyrell and Washington Counties on the Albemarle Sound and giving Gates County back to the VRA district to the west. This makes the district a whole percentage point more Democratic and given the incredibly high degree of correlation between statewide performance and senatorial performance and the fact that appointed incumbent Democrat Stan White lost to challenger Bill Cook by just 21 votes, this is a district where removing gerrymandering almost certainly changes the outcome.
SD-03/SD-04
Republicans used the VRA majority minority districts to pack in Democrats, but not in ways that are at all required. Given how this region of the state has either grown much slower than places like the Triangle or outright lost population since the 2000 census, it was at the very least just as easy to draw majority minority districts then as it was following the 2010 census. After that round of redistricting, the districts drawn that split few to no counties were upheld as VRA compliant by the Department of Justice or by the state level courts and I see no reason why the same number of districts wouldn't also be legal in essentially the same location. Therefore I've reverted these two back to what is largely the pre-2012 versions but have now split Nash County so that the blacker parts of it including Rocky Mount can go into the 3rd. Additionally, unlike in federal redistricting where court cases have forced districts to be 50%+1 black VAP where possible, the rules are a little less stringent for the state legislature while the standard for reasonable compactness is higher and as such both districts are around half black by voting age population and majority black overall.
SD-05
This district becomes much more compact by taking in the whole of Lenoir and Wayne counties rather than becoming a 3rd rural VRA district. The 2002-2010 maps did not turn this into a majority black district when it was just as easy to do so and I see no reason why it should be forced to. Under these lines, the district is no longer safe Democratic but would have seen incumbent freshman Republican Louis Pate running here rather than in the gerrymandered 7th. The district was almost exactly the same as the state in terms of downballot partisanship at D+0 and thus would have been a complete tossup. Democrats have a strong bench here though and it's quite likely that senator Donald Davis (who is black), whom Pate defeated in 2010 in a similar district but who ran in the much safer 5th district last year, would have run for a rematch here instead. While the district could definitely have gone either way, I think Davis would have very narrowly eked out a win.
However if you think that making this district a majority black one is required, that still doesn't force the drawing of the neighboring 3rd and 4th how Republicans drew them. With those two districts very similar to how I drew them in the main map, you can also get a heavily black 5th district though it isn't quite compact at all, but splits the same 5 counties plus just one more compared to the Republicans' map. The net effect is that, instead of two tossup districts, the 5th and 11th are safe for Democrats and Republicans, respectively. All three of the 3rd, 4th, and 5th here are just over majority black by total population.
SD-07
The 7th district that Republicans drew is wholesale dismantled and the previous iteration is largely restored, comprising Franklin, Granville, and Vance counties which they had split between three districts, plus a small slice of northeastern Wake. Incumbent Doug Berger, who had run in a Republican leaning successor district in 2012 and lost, would be absolutely safe here in this D+7 solidly Democratic district that even Walter Dalton won.
SD-08
The 8th district doesn't change all that radically with the biggest difference being that it drops Bladen County which it had added in 2012 and regains southern Columbus County, of which it had previously held the entirety. I've also removed the two heavily Democratic precincts from downtown Wilmington and added one outside the city proper and its suburbs along the beach further south. These changes move the district 3% to the right and make it quite safe for Republican Bill Rabon, though given the area's trend he's in no danger in his actual district.
SD-09
Just rearranging three precincts as I did in New Hanover County makes the 9th district roughly 2% more Democratic thanks to regaining parts of dark blue downtown Wilmington and dropping conservative beach areas. Freshman Republican Thom Goolsby won by just over 8% and here that margin would have narrowed to less than 5% despite a spending advantage of nearly 4 to 1. While it's probable that he still would have won, Goolsby would be significantly more vulnerable in 2014 where it wouldn't take much for him to lose in a more favorable Democratic environment.
SD-10
District 10's partisanship changes dramatically with it no longer venturing into parts of ultra Republican Johnston County and instead picking up Bladen County and northern Columbus. Though it trended strongly Republican between 2008 and 2012, the district was still more than 2% more Democratic than the state at large and is a whole 8% more Democratic than the actual version. While incumbent freshman Republican Brent Jackson, who was swept in with the 2010 wave, went unopposed in 2012, that would almost certainly not have occurred here. This district has a very deep Democratic bench and Jackson only won by 4.5% in what was a 3% more Republican open seat in 2010. Republicans in the state legislature only hold a single district that was this Democratic in 2012 where they beat an underfunded Democratic challenger in an area with a relatively weaker bench. Therefore I find it very unlikely that they would have held onto this district with the statewide Democratic ticket winning by roughly 6% here. However the district would have undoubtedly been very competitive.
SD-11
Like the 10th, this district gets very significantly more Democratic without a Republican gerrymander allowing them to split heavily Republican Johnston County. The district retains the entirety of Wilson County from its old iteration and now adds all of Greene and the southern half of Pitt, home to Greenville. Compared to the Republican version, this district is now 7% more Democratic and was D+1 downballot in 2012. While incumbent Buck Newton won a more Democratic version of this district in his initial 2010 election and only faced an unfunded token challenger in 2012, over half of this district is new to him and has a strong Democratic bench. I have a hard time making a choice if forced to over which party would win this tossup district, but it would definitely depend on candidates. If Democrats landed a strong, decently funded candidate, they probably narrowly win, if they get someone mediocre like they did in the similarly light blue SD-19 in 2012, they probably don't.
SD-12
The actual 12th district was surprisingly competitive given its R+5.5 partisanship and saw one of the narrowest Republican victories in 2012, but here that's definitely not the case. This district overwhelmingly reverts back to its 2010 configuration with the entirety of Johnston County and now adds southern Nash. At R+12 it's now one of the most Republican districts in the eastern half of the state and the seat left open by then incumbent David Rouzer for a losing congressional bid would easily send a Republican replacement to Raleigh.
SD-15
Republicans packed Democrats in Wake County into just 2 districts and sent this district into strongly Republican north Raleigh. Here I've moved it back north of the I-440 inner loop to be comprised entirely of the fast growing but heavily white and thus Republican leaning suburban areas north of central Raleigh. Subsequently, the district moves 2% in the Democrats direction and now voted very narrowly for their statewide ticket on average, but unlike many of the districts discussed above, Republican incumbent Neal Hunt has been in office since 2005 and is fairly entrenched here. Seeing as he won by about 11.5% he most likely would have been held under 10% now by his 2012 challenger, but wouldn't have ever been in danger of losing. Though Democrats wouldn't have won it in 2012, at just barely more Republican than the state and trending Democratic at a strong pace, Hunt would become a top Democratic target in 2014 and would be likely to lose in the event of a pro-Democratic wave as recent polling has suggested.
SD-17
Conversely from the 15th, the Republican leaning 17th in southern Wake County becomes 2% more Republican than the gerrymandered version by losing most of its portion of the town (suburb) of Cary. Republican Tamara Barringer held this open seat in a competitive race in 2012 and won by a little under 10% while here that widens to just over 10%. Though the actual 17th district is one that Democrats absolutely will have to win to take the state senate, here it's a little less important. Still, as with much of Wake County the 17th is trending Democratic and though it is now R+5 downballot, it would likely still see a heavily contested race in 2014 that could potentially fall to the Democrats in a wave election.
SD-18
This Republican won district is entirely dismantled and resurrected as a safe Democratic district in Cary and southern Durham. Though it has relatively little overlap with the 2010 version of the 18th, it is possible but unlikely that then incumbent Bob Atwater could have sought reelection here instead of the neighboring 24th where much of his old district and base went. Regardless, the 18th is a lock for Democrats at D+11.
SD-19
The 19th was one of the more grotesquely gerrymandered Republican districts even though it ended up being slightly Democratic leaning anyway. Still, freshman incumbent Wesley Meredith faced an underfunded Democratic challenger whom he outspent 8:1 in 2012 and thus held on by a modest 6%. Now the district becomes more similar to its old version with Cumberland County minus its northwest quadrant and much of Fayetteville and now adds southern and eastern Harnett to the north in place of Bladen County. Despite becoming 4% more Republican, the district has a stronger Democratic base and now includes the home of real life 12th district Democratic nominee attorney Brad Salmon. Salmon lost by just 2 points in a 4% more Republican district in 2012 despite being heavily outspent and if he were the Democratic nominee rather than actual 19th district nominee George Tatum, Meredith would have had a much tougher time winning reelection. While the district's partisanship now narrowly favors Republicans and Meredith would have likely still had a strong spending advantage, this race would be much more competitive if not a true tossup. As in 2014 where the real district is Democrats' top pick up opportunity, this district would be almost certain to fall to them with a favorable political environment.
SD-20
Though the change in partisanship of this safe Democratic district is inconsequential, the geography and demographics of the district change considerably. Rather than venture into Granville County to pack in Democrats and become majority black, the district retains its pre-2012 make up by containing the vast majority of the city of Durham. Given how much less racially polarized this area is than the rural districts in the east such as the 3rd and 4th, I do not believe that it is required to be majority black by the VRA. The district is now just plurality black at 43% and would have easily sent incumbent Floyd McKissick, whose districts proposal heavily influenced my own, back to Raleigh for another term.
SD-22
Balancing out the 18th is the 22nd which Republicans drew as a Democratic vote sink containing the rest of Durham along with Caswell and Person Counties. Here it's dismantled and the old 22nd is largely restored with all of Moore County, western Harnett, and now all of Lee County and becomes strongly Republican at R+8. 10 year incumbent Republican Harris Blake might not have retired and would have been a shoo-in for another term, but even if he retired anyway as seems likely, Robin Rabin who won the actual 12th district probably would have held this one instead for the Republicans.
SD-24
This district changes drastically and is the only one we see that would have drawn in 2 incumbents of different parties, setting up a D vs R incumbent showdown in November. Freshman Republican Rick Gunn only won by 6% in a Republican leaning district in 2010 in a rematch against the then freshman Democratic incumbent. 2/3rds of the new district, the vast majority of Alamance County, comes from Gunn's old 24th district while just the remaining 1/3rd comes from Chatham County, the base of then Democratic incumbent Bob Atwater. However, Atwater had been in office since 2005 and though he previously represented a relatively safe Democratic district, this new one is just under R+2 and is a few points more Democratic than the one Gunn previously represented. This district would assuredly be a tossup and really could go either way, but Gunn probably has slightly over 50/50 odds of winning due to partisanship and geography. In a Democratic wave though, this district is almost a sure flip.
SD-25
The 25th district in central NC along the South Carolina border was the only Romney won, R+ average district that Democrats won in 2012. Here however, it drops the dark red and monolithically white rural precincts in Rowan County and adds nearer, more diverse, and much less Republican Montgomery County. Though the actual district was slightly Republican leaning, Democrat Gene McLaurin held onto the open seat with a relatively comfortable 6% win. With this district moving a whole 4% towards the Democrats, McLaurin would have held it with ease.
SD-27
Reversing the Republican gerrymander which packed the most Democratic areas of Guilford County into the majority black 28th district, the old 27th district is revived. Here it drops all of rural Guilford and High Point and contracts into just the city of Greensboro. Going from heavily republican to heavily Democratic, this district would have easily reelected then incumbent senator Don Vaughan who retired rather than face certain loss in 2012. The neighboring 28th is still roughly half black and now takes in the entire city of High Point plus the heavily minority southern and eastern sections of Greensboro.
SD-47
The last district where the partisan change is worth discussing is what I have now numbered the 47th, replacing what Republicans drew as the 41st in their map. That district stretched around the entire city of Charlotte to take in as much of the heavily white Republican suburbs, but here it is a similarly new district in the northern tip of Mecklenburg County and northwest Charlotte and replacing an eliminated Republican district in western NC. Though Democrats didn't contest the 41st last year, that district was strongly Republican and this one was not only D+3 downballot but also saw Obama win by nearly 10% in 2012. Republicans maintain a strong base in Mecklenburg County, but it's very difficult to see them winning this open seat.
The remaining district changes were largely geographic in nature or were just too one sided for the partisan change to have mattered. All in all, the Democrats are in a much better position in the senate elections and given the districts' partisanship, plausible candidates, and the strong degree of correlation between statewide and local candidate performance, here is how I feel rating the competitiveness of the 50 districts:
And with those ratings and all of the above in mind, here's what the difference might have been between the map I've come up with and the Republican one used in the actual elections:
Guaranteed or very likely Democratic flips are in vivid blue while possible flips are in light blue and possible Republican flips are in pink, giving us a total change of Democrats +4 to +7 and leaving them with 21-24 out of 50 districts, needing 26 for control. If forced to pick out of those three that could go either way I think Democrats would have won the Dem leaning 5th and 11th and Republicans the GOP leaning 24th, leaving the chamber divided 27R-23D. While under the actual map, Democrats seem to have a ceiling of 24 districts, here they could easily expand up to 28 in a favorable statewide political climate in 2014.
Thus, you can see that Republican gerrymandering played a significant role in securing the state senate for them and almost certainly was what allowed them to keep their veto proof majority. While Democrats likely would have fallen just a few seats short, I've elaborated above on how they would easily have a path to the majority if 2014 ends up anywhere near as favorable as recent polling has suggested with a Democratic generic ballot lead in mid to high single digits. This is drastically improved from the actual legislative districts where I've previously detailed just how difficult it will be for the Democrats to gain the 9 seats they need in 2014.
If you're interested in my data courtesy of DKE's jeffmd and the North Carolina General Assembly and State Board of Elections, you can find the current legislative districts along with the 2012 precinct file here and the data and DRA map file for my alternate senate districts here.