Welcome to Part 2 of my look at the NC State Legislature with an analysis of the House. You can find Part 1 on the Senate here.
Following the 2012 elections, North Carolina saw its first Republican-led one party government since Reconstruction. Thanks to their control over the redistricting process, Republicans were able to gerrymander the state legislature so effectively that they now have veto proof majorities in both chambers, not that newly elected governor Pat McCrory would veto their agenda. Despite electing McCrory, North Carolina didn't see some massive Republican landslide statewide, but rather gerrymandering, or the drawing of political districts to strongly favor a particular party, enabled them to win sizable majorities. Six months into Republican-led government, Pat McCrory has seen his popularity tumble to where his approval rating is no longer positive. The Republicans in the legislature are horribly unpopular and while the Democrats don't have that great of a public image, the disparity is enough that the generic ballot asking whether poll respondents favor one party over the other in general has consistently found Democratic leads.
However, given how extreme this gerrymandering is, it is still a daunting task for Democrats to win either chamber of the legislature despite how unpopular Republicans get. So let's take a look at the 2012 elections and why the results give us ample reason to believe it will take a massive statewide lead for the Democrats for them to take either chamber, now focusing on the House of Representatives.
For reference these are the 120 state senate districts. You can find more detailed maps here.
Democrats had long controlled the North Carolina General Assembly as many conservative Democratic parties had done for generations in the South. However, like many legislatures including the US Congress, NC Republicans made massive gains in the 2010 landslide and wrested control from the Democrats for the first time since Reconstruction. Since 2011 was a redistricting year and then governor Bev Perdue had no power to veto, the General Assembly's Republicans passed an aggressive gerrymander of the legislature. The previous map used in the 2004-2010 elections was drawn as a compromise between the then Democratic Senate and Republican House and was by most regards a fair, non-partisan map. After the 2010 elections, Republicans held a 68-52 edge in the chamber with that map.
Looking at the 2012 elections you can see that, unlike the state senate where Republican gerrymandering largely protected their 2010 gains, they were actually able to produce a considerable gain of seats and netted 9 more after some of their seats were eaten by redistricting. Thus, their majority expanded to 77 to 43 and is now veto-proof. While some odd looking districts are required by the Voting Rights Act and the need to maintain majority minority districts, others such as those in Mecklenburg and Wake counties are clearly gerrymandered to give Republicans an advantage. Also unlike the senate map, the House was able to be more aggressive because smaller districts make it easier to justify splitting counties between multiple seats. The North Carolina constitution explicitly prohibits what it deems "excessive county splits" and the previous Democratic gerrymander from 2002 was struck down in court and never used for that very reason. Let's look at the 2012 districts by the numbers.
Here I've given the district demographics from the 2010 census, the two-party only state House vote (all election data will exclude 3rd parties), the 2012 presidential vote, the 2012 PVI (Partisan Voter Index), which measures the percentage point difference between the district vote share and national vote share, the 2012 average of the 9 statewide executive races such as governor, and the average PVI with the state as a whole being evenly balanced (though 50.6% Democratic). Looking at the demographics of the state, we can see that 25 districts are over 40% Black while one is majority Native American and an additional one is majority minority and the remaining 93 are majority white.
In this map I've colored the districts by their 2012 presidential vote. Statewide, Mitt Romney won by 2% yet it's fairly clear that he won the vast majority of districts here, 80 in total for an even 2/3rds.
When we look at the graphical distribution of the seats by their presidential numbers, you can really see the strong Republican advantage in the map. Nationally, North Carolina was R+3 compared to the nation meaning that Obama received 3% less of the vote and Romney 3% more than they did nationwide. Thus I've set R+3 as being evenly divided between the parties with those districts to the right increasingly stronger for Republicans and vice versa for Democrats on the left. The black line represents the district that would flip control to the Democrats, the 61st least Democratic House district. The overwhelming majority of the districts favor Republicans and nearly half do by a very significant margin. The 61st district in the chamber is R+9 and gave Romney 57% of the vote, or 6% to the right of the state. Under the old map, the 61st district was just R+6.6.
However, like many places in the south in particular, local candidates don't always perform in ways we would expect based solely off of the presidential election. Concurrent with the presidential election, North Carolina held elections for 9 statewide executive positions (Attorney General was uncontested by Republicans). Republicans won the positions of Governor, Lieutenant Governor, Agriculture Commissioner, and Labor Commissioner while Democrats won the positions of Secretary of State, Treasurer, Auditor, Schools Superintendent, and Insurance Commissioner. Overall the average of the 9 races statewide went 50.6% for the Democrats, even closer than the presidential election. Here's what the 120 House districts look like when you compare the district by district average to the statewide average:
Notice that I have narrowed the scale, but that it still shows that Republicans have a big advantage with a total of 78 out of 120 being more Republican on average than the state. Additionally, very few districts are even close to the statewide average as they are designed to pack as many Democrats into as few districts as possible while spreading Republicans out into many that are strongly Republican but not overly so. Only 1/6th of the districts were even within 5% of the statewide average either way.
As with the presidential performance, the graph just really hammers home how Republican the map is. The majority of districts strongly favor Republicans and the 61st district is R+6 compared to the state. Under the previous map it was significantly lower at just R+2.1.
Now that we've seen the districts by all 10 federal and statewide executive elections, let's take a look at the House of Representatives performances themselves. Races that were uncontested by one of the two major parties will appear as the darkest shade.
Astoundingly, a total of 55 races weren't contested by both parties leaving just 65 with a Democratic and Republican candidate. Unlike in the Senate where there were nearly twice as many uncontested Republicans as Democrats, the split was basically equal in the House with 28 Republicans not facing a Democratic opponent and 27 Democrats not facing a Republican. However, these Democratic districts were much more Democratic than the Republican ones were Republican. When we take into account these seats based upon those that were contested and the average/presidential results, the Republican popular vote win of 3% actually becomes a Democratic popular vote win of 1%. Given how close this popular vote was and how Democrats are geographically more concentrated than Republicans, it is unlikely that gerrymandering was the single cause of Republicans holding the chamber, but it absolutely contributed to their ability to turn a 1% loss into 64% of the House seats.
Before we can begin looking at what 2014 might hold there are several more sets of data that require our attention. For one thing, you might have noticed that the parties' district by district performance didn't nicely match up with presidential or even the average in certain districts.
Here I've illustrated for reference the difference between Obama's vote share and the statewide average. In all of NC Obama ran 1.6% behind the rest of the ticket as a whole, but he ran ahead or even with it in all of the more urban and suburban districts in the Triangle, Triad, and Charlotte and even Wilmington while he ran behind it the most in rural eastern North Carolina. While that region as part of the national Black Belt region has a large rural black population, it has for generations been the bastion of white conservative Democrats, a demographic who have been voting Republican for president for a several cycles now and it is no surprise that Obama would underperform the ticket there.
In this map I've subtracted Obama's vote share from the Democratic candidate with uncontested seats in gray. What we generally see is the Democrat running far ahead of Obama in many of the more rural districts while running behind in some of the more suburban or heavily white urban ones. Incumbency also plays a role as the few rural districts where Republicans ran ahead of Romney, they had incumbents for the most part. Among all the contested seats, the r-squared value for the comparison of Obama and House Democrats was .92 which is fairly high.
Taking a look at the Democratic candidate compared to the statewide candidates' average, we see a map that has a little less disparity. Democrats still run ahead in the some of rural districts, partly because they're more conservative than the statewide ticket while Republicans tend to run ahead in the more suburban ones due to incumbency both their own and that of the majority Democratic incumbent statewide officeholders. However, many more districts see the Democrat doing less than a point better or worse than average. Overall, the r-squared was .96 which is very high. As such, I prefer using the average as a means of gauging a district's relative competitiveness though the presidential numbers aren't without merit.
You can see how clear this relationship was though a few districts stick out where the winner or loser performed above what we might expect on paper. For every point decrease in the average PVI, the Democratic candidate saw a 1.1% decrease in vote share on average. Now that we've analyzed the districts and investigated what data can serve as a useful indicator of performance, let's take a look at how 2014 might shape up.
One logical place to look in our search for which districts might change hands would be those that were the closest in the preceding election. However, not all districts have the same fundamentals and just because a race was close one year doesn't necessarily mean it will be close the next. Let's take a look at some of these changing fundamentals and how they will impact 2014.
This map looks at the change in the presidential PVI or the relative change in the distance from the national result among each of the districts from 2008 to 2012. The state as a whole trended about 1 point Democratic and here we see that those districts in places like Charlotte, Raleigh, or northeastern and southeastern NC, nearly all of which have significant minority populations, trended Democratic the most. Those that were the whitest such as in western NC trended Republican the most but surprisingly so did a few white majority districts in places like Mecklenburg and Wake Counties.
For the most part, this trend is a net negative for the Democrats and positive for the Republicans and while a few districts might reverse such as those in Mecklenburg or Wake, it is highly unlikely that the rural white majority districts won't continue to trend Republican as they have done for several cycles now.
Above though we saw that the presidential performance was a slightly inferior metric than the statewide average in trying to predict senatorial performance. This map shows the districts by the relative change in their statewide average PVI, or the change in the distance between the district average and the state average over 2008-2012. Again we see all of the urban/suburban districts trending Democratic aside from Wilmington while even here the black majority rural ones are neutral at worst for the Democrats. Even more pronounced though is the Republican trend of nearly all of the rural white majority districts, particularly in the southeast and the mountains.
Even more so than the presidential trend, this map is a very clear net positive for Republicans and shows the ingenuity of their gerrymander. While there are several districts that saw a significant trend towards Democrats, the vast majority of them are currently held by the Democrats and are designed to pack in as many of their voters as possible in places like Charlotte, Raleigh, etc. Unlike in the presidential trend, we can pretty clearly see that just about all of the suburban Charlotte and Raleigh seats are trending Democratic, but regardless, even if they won every remaining seat in Wake and Mecklenburg counties, Democrats would gain just 12 out of the 18 they need to win the chamber.
The final trend map doesn't look at relative party strength but instead looks at relative rates of ticket splitting. To make an apples to apples comparison, I've averaged the downballot statewide office vote shares and compared them to the presidential vote share in both 2008 and 2012. Districts where that difference was wider in 2012 than in 2008 are in blue while those where that difference is lower, and thus saw more of an alignment over that period are in red. As you can see, the vast majority of districts saw a decrease in ticket splitting between 2008 and 2012.
Again, this is a major net positive for the Republican party. This is because the districts where Democrats rely on ticket splitting the most, those Romney-won rural white majority districts, saw ticket splitting fall the most. This is why having a Dem-friendly or neutral presidential PVI shift in that preceding map was not beneficial to the Democrats in rural districts. Even if voters are becoming more open to Democratic presidential nominees, they still don't make up a majority of the electorate and the voters needed to win these seats are becoming considerably less open to local Democrats. This is somewhat counterbalanced by the more suburban districts, but those among them that are potentially competitive are fewer in number than the rural ones which are presently or were until very recently competitive.
Okay, so at this point you've probably seen enough maps and want to know what you might be able to expect from the 2014 elections. Let's take a look at polling and while I normally would caution against relying on a single polling firm, the Raleigh based Public Policy Polling was one of the most accurate pollsters across the last couple of election cycles and since they're based in North Carolina, they poll statewide once a month and give us a great trend line.
Here we have 4 different data points measuring the popularity of Pat McCrory, Legislative Democrats and Republicans, and the spread between the parties on the generic ballot test.
Starting from the top with the actual gubernatorial results and the uncontested-race adjusted state senate vote, you can see four trends that are quite clear. The first is that Pat McCrory, like many first term incumbents, saw a honeymoon period where his approval was far higher than his disapproval; however that is at an end. His approval has been stagnant for several months while disapproval has been rising. The generic ballot has also favored the Democrats every single time and since March by a significant amount. Next, legislative Democrats are unpopular but not terribly so. Overall they've floated around high single digits in net disapproval. However that's significantly better than Republicans who have consistently found themselves with double digit net disapproval ratings which on average are in the mid to upper teens. Given that finding it's no surprise that Democrats possess a generic ballot polling lead.
While this last poll might be an outlier in terms of Democratic fortunes (though probably not), what it and many previous polls show is that the legislation the Republican General Assembly has put forth is unpopular. Really unpopular. Opposition to the Republican budget runs nearly 3-1. Opposition to their abortion bill outweighs support for it by double digits. Opposition to their cutting off unemployment benefits for tens of thousands of North Carolinians runs 2-1. Regardless of whether the polling is a few points too rosy for Democrats those numbers are insurmountable. If Republicans continue on their current legislative agenda and all signs point to that being the case, they will remain unpopular and Pat McCrory, their lone shining star, will be tarnished.
Based on those factors and how we can expect them to unfold over the next 16 months, it is incredibly likely that the 2014 election environment will favor the Democratic party. If the generic ballot lead winds up being remotely close to D+9, Democrats will almost certainly gain several seats, but even if it is lower than that a gain is likely and Republicans would really need a much more positive environment to gain seats.
Still where might Republicans gain districts if they are to try? Unlike the Senate where Democrats hit rock bottom, there are a few districts that might switch to Republicans if the statewide environment were to become more favorable to them and are worth taking a look at.
The 6th district in eastern NC is by far the most Republican district that Democrats hold, yet Paul Tine won it in an upset as an open seat in 2012 despite Romney winning by over 16% and the Republican statewide ticket winning by an average of 8%. The district is R+5 on average and trended fairly strongly Republican in 2012 so it would not be surprising to see Republicans win it at some point. The 2nd district Republicans might gain is the 22nd in the southeast where incumbent William Brisson actually went uncontested, yet he's only been in office since 2006. The district is slightly more Republican than the state and trending more so while Romney carried it easily, but Brisson is a formidable incumbent. The final district Republicans have a decent shot at flipping is the 119th in far western NC. It's quite swingy but trending strongly Republican, however Democrat Joe Sam Queen held it as an open seat in 2012. Overall, none of these three is likely to flip without a much less hostile environment for the GOP, which leaves us with the 18 seats Democrats would need to take.
The first tier of seats should absolutely fall to the Democrats if the repeated D+4 or greater generic ballots are to be believed. At the top of the list is one I'm astounded that Republicans won in the first place, district 92 in western Mecklenburg County. Not only was it an open seat in 2012 that Republican Charles Jeter won, but it's also D+1 on average and amazingly D+1 federally meaning Obama won it by 6%. I don't know of a single other state legislative seat in the South that is both D+ and elects its members in presidential years. Anyway, this district is trending sharply Democratic and should continue to do so. Jeter only won by 3% and would be toast in a wave election and even in a neutral environment would have a tough reelection bid. After that district there's just one other Republican held House seat won by a republican, the 41st in southwestern Wake County where freshman Rep. Thomas Murray won reelection by just 3.5%. Like most of fast growing suburban North Carolina, this district is zooming to the left and was very evenly divided between the parties in 2012. Though that's it for Obama districts, there are two others that are barely Republican leaning and saw the Republicans win narrowly in 2012. The first is the 49th which Jim Fulghum held as an open seat last year. As befits such a district, this north Raleigh district is trending strongly Democratic and was only barely more Republican than the state in 2012. Fulghum won by 8%, but the district dynamics are just going to keep getting less favorable. The final district in this set is the 115th in eastern Buncombe County (Asheville). Republicans gerrymandered this district to drop Asheville and thus it was open in 2012, but it's very light red unlike many other prospects. Obama lost it by just over 3% and the Democratic ticket by 1% making it R+1 on average. Also unlike many western NC districts it is not getting any worse for the Democratic party and if anything is slightly improving.
The 2nd tier of districts won't be nearly as easy for the Democrats to win but is also comprised of districts that will be essential to winning a majority and should be within reach under a strongly Democratic environment. Up first is the 35th in northeastern Wake County. Republicans gerrymandered this district to favor them and just narrowly won the open seat in by 2% in 2012 with Chris Malone. The district is trending Democratic though not as strongly as some others in Wake County and at R+4 on average isn't exactly easy turf, but it nonetheless is required for a majority. District 51 in Lee and Harnett counties is similarly red at R+5 on average but isn't trending really one way or the other. Freshman Republican Michael Stone won by just 4% and given his relatively short tenure should be an easier target than others in this category. Additionally there is the 8th district in eastern NC where Susan Martin won the open seat by a surprisingly large margin for what was just an R+2 district on average. Unlike many white majority rural districts it isn't trending Republican and should be a definite target for Democrats. Just nearby is the 9th district which Republicans were able to gerrymander so as to defeat longtime incumbent Democrat Marian McLawhorn by just 3%. That seat is R+4 on average but is trending modestly Republican. After that is likely the 36th in suburban southern Wake County where Nelson Dollar has been in office since 2004. However this seat is trending strongly Democratic to where it was just R+4 on average in 2012 and Dollar won by only 10% and barely ran ahead of Mitt Romney. Similarly is the 40th in northwestern Wake County where Marilyn Avila first won in 2006 and is a similar R+4 on average but trending even more Democratic. Avila won by a wider 12.5% but the district dynamics should make her vulnerable. Next is the 45th in Cumberland County which despite being open in 2012 saw John Szoka win by a wide margin. While the district was just R+3 it is trending modestly Republican but should be winnable. After that is the 93rd district in Ashe and Watauga counties in northwestern NC. Freshman Republican Jonathan Jordan won by just 3% in 2012 but like many Appalachian districts it is trending Republican and at R+5 on average is fairly Republican leaning. Finally there is the 118th district in far western NC north and west of Buncombe County. Five term Democratic incumbent Ray Rapp narrowly lost due to the Republican gerrymander, but this district was R+5 on average in 2012 and trending very strongly Republican. The district is still one of the less difficult for the Democrats to win on their way to 61, but it won't be an easy one.
In this last tier are districts that quite favorable to Republicans or have fairly entrenched incumbents but are probably still required for a Democratic majority. As such I only see Democrats being able to gain the previously mentioned 13 seats under present conditions. Regardless though, they would need 5 more and one of the more appealing yet challenging districts in this group is the 37th in southern Wake County which is R+6 downballot but trending modestly Democratic. What makes this district stand out is that it is home 5 term Rep. and current House Majority Leader Paul Stam. With the General Assembly and legislative Republicans and their policy agenda about as popular as a colonoscopy, Stam should be a huge target for the Democrats as one of their leaders. However he is well entrenched here and would have easy access to campaign funds, but as we've seen at the federal level in poll after poll, leaders of an unpopular party can quickly become unpopular no matter how formidable a candidate they might be. Another district would be the 116th west of Asheville which was gerrymandered to protect freshman incumbent Tim Moffitt. He won by a wide 12.5% margin in 2012, but the district is one of the least red left this far down the list at R+6 on average and isn't really trending one way or the other. Somewhat similar is the 53rd in Harnett County which at just R+5 on average and neutral in trend should be a decent target. However incumbent David Lewis has been in office since 2002 and won by 13% in 2012 and is quite entrenched. The last two districts aren't any more friendly to Democrats. The 1st district in northeastern NC was gerrymandered to defeat the Democratic incumbent who retired instead. Bob Steinberg won the open seat by a comfortable margin and Romney crushed Obama, however this district was R+6 on average but importantly trending modestly Republican. Finally there is the 88th which was yet another gerrymandered pick up. Long time Democratic incumbent Sandra Alexander lost by about 10% in this Dem trending district, but it is an imposing R+7 on average and would likely take a huge wave to dislodge.
Just like in the Senate, Democrats are almost guaranteed to gain some seats in the House of Representatives, but the closer they get to the number needed for a majority it becomes incredibly difficult for them to win thanks to the ingenuity of the Republican gerrymander.
To conclude, I'd just like to drive home a few key points:
- Democrats likely won the House popular vote in 2012
- Republicans' ironclad grip on the legislature is solely due to gerrymandering
- Democrats will need a massive level of public support to win under this map
- Republicans are in a great position to keep the legislature for possibly over a
decade because the strongest Democratic growth is in Democratic held districts
- Democrats are almost certain to gain seats, but probably not the 18 they need
Sometime next year when we're closer to the election I'll have an updated look at the state legislative elections once we have a good idea of which candidates are running and where the parties plan on competing. My next diary in this series will take a look at alternative non-partisan maps to investigate what the impact of gerrymandering was in 2012 and what it might be in 2014.
If you're interested in any of my data it is publicly available here. All 2008 election data was from the General Assembly itself while the 2012 data was courtesy DKE's jeffmd.