My Proposal for how Pennsylvania should Redraw its Congressional Districts

In late January, the Pennsylvania Supreme Court ruled that the state’s Republican-drawn congressional district map violated the state constitution’s guarantee of free and equal elections. That map, which was brutally gerrymandered, consistently sent a congressional delegation of 13 Republicans and 5 Democrats to Washington despite Pennsylvania generally being a fairly evenly divided state politically. The Court said that the new map must minimize county and municipality splits and that districts must be geographically compact.

With that in mind, and with my longstanding interest in both redistricting and the political geography of Pennsylvania, I have drawn a map that satisfies all of the Court’s criteria. The map was specifically drawn to not favor any political party, and it splits only 10 counties statewide (three of which, Philadelphia, Allegheny, and Montgomery, must be split due to their having a larger population than a congressional district). The map was also drawn without regard to the hometowns of any members of Congress. I drew the map using Dave’s Redistricting App, which can be found here.

PA 18 dis 2016 legal no gerry whole state

Here are the 2016 and 2008 election results for my proposed districts:

District Clinton Trump Castle Stein Johnson Total Clinton % Trump % Obama % McCain % PVI
District 1 170,831 177,402 1,126 2,467 8,453 360,279 47.42% 49.24% 52.40% 46.50% R+1.3
District 2 224,358 135,490 1,025 3,239 9,572 373,684 60.04% 36.26% 60.30% 38.80% D+9.8
District 3 116,196 213,021 1,484 2,199 9,360 342,260 33.95% 62.24% 43.20% 55.60% R+13.4
District 4 101,228 205,085 1,133 1,644 7,310 316,400 31.99% 64.82% 42.50% 56.30% R+16.2
District 5 91,173 213,504 1,450 2,139 7,501 315,767 28.87% 67.61% 41.80% 56.80% R+18.8
District 6 105,227 185,277 1,374 2,401 8,599 302,878 34.74% 61.17% 49.80% 48.80% R+10.5
District 7 139,003 158,272 689 2,623 5,256 305,843 45.45% 51.75% 58.40% 40.60% D+0.9
District 8 85,884 205,768 1,375 2,421 7,813 303,261 28.32% 67.85% 41.90% 56.70% R+17.5
District 9 102,890 214,097 1,436 2,511 9,712 330,646 31.12% 64.75% 40.30% 58.60% R+16.8
District 10 119,194 191,465 1,926 2,654 10,795 326,034 36.56% 58.73% 43.30% 55.80% R+12.5
District 11 136,140 193,095 1,567 2,792 9,680 343,274 39.66% 56.25% 47.40% 51.40% R+8.1
District 12 157,396 161,449 1,139 3,115 8,600 331,699 47.45% 48.67% 55.80% 42.80% R+0.6
District 13 185,388 155,020 1,149 3,247 10,210 355,014 52.22% 43.67% 56.00% 43.00% D+1.7
District 14 240,272 130,482 1,105 3,220 7,594 382,673 62.79% 34.10% 62.70% 36.40% D+12.2
District 15 183,158 178,499 1,287 3,382 9,026 375,352 48.80% 47.56% 54.00% 44.90% R+0.6
District 16 216,140 148,728 1,157 3,301 9,821 379,147 57.01% 39.23% 58.80% 40.30% D+6.2
District 17 228,795 59,085 566 2,812 3,257 294,515 77.69% 20.06% 76.60% 22.70% D+29.2
District 18 321,579 34,759 401 3,492 3,313 363,544 88.46% 9.56% 90.40% 9.10% D+38.9

District 1 takes in the Southwestern Pennsylvania counties of Greene and Washington as well as a wide swath of Pittsburgh’s southern and eastern suburbs. This area, with the exception of a half-dozen or so wealthy suburbs of Pittsburgh, is mostly working-class and used to be heavily Democratic. However, much of the district, particularly the rural areas, has trended heavily Republican recently, even as the wealthy suburbs have trended Democratic. Barack Obama won this district by 6 percent in 2008, but it flipped to Trump by a slim 2-percent margin. This district would be a swing district, with both parties having a fairly good shot at winning it. Republicans would need to keep up Trump’s margins in Washington and Greene while improving on Trump’s lackluster performance in those wealthy suburbs of Pittsburgh, while Democrats would need to keep some of their ancestral strength in the rural areas, perform strongly in the wealthy suburbs, and get out the vote in the heavily Democratic inner eastern suburbs to win this district.

The district forms a community of interest by placing Pittsburgh’s working-class southeastern suburbs with other working-class towns in Washington and Greene counties. Those areas really don’t go well with the more middle- and upper-class northern suburbs of Pittsburgh. The wealthy southern suburbs of Pittsburgh are in this district for compactness purposes – if they were in the 2nd, then the 1st would be underpopulated, and any other configuration of the 1st would make the 3rd and 4th districts substantially less compact. But maybe I’m getting ahead of myself.

District 2 includes the entire city of Pittsburgh and also takes in most of its western, northern, and northeastern suburbs in Allegheny County. The district is entirely contained in that county, and the suburbs that it includes are generally more middle-class than those in the 1st, which helps the 2nd form a community of interest. This district is solidly Democratic (the only solidly Democratic district in the state outside of SEPA), having voted for Hillary Clinton by a 24-point margin. Many of the suburban towns here lean Republican but are slowly trending Democratic, while Pittsburgh itself is heavily and reliably Democratic at all levels.

District 3 is a compact district north of Pittsburgh that contains the three working-class counties of Beaver, Lawrence, and Mercer (those three counties constitute a community of interest) as well as the remainder of Allegheny County, all of Butler and Venango, and a portion of Crawford to equalize the population. This district is mostly working-class except for the northern suburbs of Pittsburgh that are located here – the 2nd already had enough people and would have been overpopulated if it had tried to include them. While Democrats used to have considerable electoral strength in Beaver, Lawrence, and Mercer, this district is now heavily Republican, voting for Trump by a 62-34 margin, and Democrats would likely not seriously contest this district.

District 4 is a compact district east of Pittsburgh that contains six whole counties – Fayette, Westmoreland, Indiana, Armstrong, Clarion, and Forest. All of those counties are largely working-class, and (at least in 2016) voted heavily Republican. Westmoreland contains just over half of the district’s population, and dominates the district politically. While Fayette used to be solidly working-class Democratic, it has since shifted substantially to the right, making it more in line with the other counties in this district. The district voted for Trump by a 65-32 margin, and Democrats would not seriously contest it.

District 5 is located in rural south-central Pennsylvania. Like District 4, it is also a stronghold of the white working-class. The areas here that used to be Democratic (Cambria County) have swung heavily Republican recently, and the areas that have been Republican for a long time (most of the rest of the district) have become even more Republican. The only place here that does not fit the community of interest is the State College area, which is well-educated and generally votes for Democrats, however considering the shape of the 6th, there wasn’t really much of an alternative to putting State College here, and the 5th is still compact. State College doesn’t really match demographically any of the areas surrounding it (it is a proverbial “island of blue in a sea of red”), so this configuration is the best that can be done. The only other county that is split here is Snyder, and that split exists solely for population equality purposes.

Despite the presence of State College, this district as a whole is extremely Republican. It is the reddest district on the map based on PVI, and is just barely Trump’s second-best district (he did slightly better in the 8th). Democrats will not seriously contest this.

District 6 rounds out the Western Pennsylvania districts, stretching from Erie to Lock Haven and picking up Meadville, Warren, Bradford, and St. Marys along the way. The only counties that are split here are Crawford (split with the 3rd) and Centre (split with the 5th). With the exception of Erie, this district is rural, working-class, and heavily Republican. This district actually barely voted for Obama in 2008 due to his strong win in Erie and smaller-than-normal losses in the rest of the district. However, the entire district, including Erie, swung heavily to the right in 2016, Trump won it by a 61-35 margin, and its PVI is R+10.5. If Democrats were feeling particularly ambitious, they might try to contest this, but they would have a very hard time winning here due to the prevailing trends – they would need to win Erie County in a landslide, get out the vote in college towns like Meadville and Lock Haven, and try to limit their losses in the working-class rural areas, all simultaneously. That’s a very tall order.

District 7 is based in Scranton and Wilkes-Barre, and also includes the counties of Monroe and Carbon. This district is a community of interest by putting pretty much all of the urban and suburban areas of NEPA into one district, allowing the 8th to be almost entirely rural. While Carbon County is also mostly rural, it is included in the 7th for compactness purposes. The split of Luzerne on this map is pretty much inevitable – the built-up Wilkes-Barre/Pittston area of the county is a different community of interest from the rural areas. While this district used to be solidly Democratic, voting for Obama by 18 points in 2008, it swung heavily Republican in 2016, giving Trump a 6-point victory. However, it still has a Democratic PVI (though just barely), and Democrats still have a lot of downballot strength in Lackawanna and Luzerne. Thus, this district would slightly favor Democrats, but Republicans would still have a chance at winning this if they can keep Trump’s coalition together.

District 8 contains rural Northeastern Pennsylvania. Its largest city is Williamsport, population 29,000, and it is uniformly Republican and mostly working-class. It stretches from anthracite coal mining areas in southern Northumberland County, to the fracking hubs of Bradford and Susquehanna, to the NYC exurbs of Pike County. It is actually somewhat similar to the 2000s-era PA-10, though substantially neater. Its only two county splits are Luzerne (with the 7th) and Snyder (with the 5th). This district is the second-reddest on my proposed map by PVI, and was Trump’s best district in 2016. Whatever Democratic strength existed here a decade ago has since collapsed, and Democrats won’t seriously contest this district.

District 9 is a very neat district in south-central Pennsylvania that contains three whole counties – York, Adams, and Franklin – and part of a fourth, Cumberland. All of those counties are solidly Republican at all levels, and they fit nicely together as a community of interest of semi-rural areas dotted with small and mid-sized towns. The district voted 65-31 for Trump, and Democrats wouldn’t seriously contest this district.

District 10 contains the counties of Lancaster and Lebanon plus a small slice of Chester County. Lancaster and Lebanon are similar because they both have a mid-sized city at the center that is generally Democratic, and extremely Republican rural areas that surround that city. They both have heavy Amish influence. The slice of Chester that is added consists of the more rural, agricultural areas of the county that haven’t yet suburbanized like the remainder of the county has, thus making that area more similar to the adjacent areas of Lancaster County than the heavily suburbanized and wealthy eastern part of Chester County. This district is solidly Republican, with Lancaster County having one of the longest streaks of being represented by Republicans in the House of any county in the country. Democrats have contested the current version of this district since it includes Reading, but this version voted 59-37 for Trump and is too Republican for any Democrat to win.

District 11 contains the Harrisburg area, and then goes east to pick up heavily-WWC Schuylkill County and a slice of Berks County. This is somewhat of a leftovers district, since neither the Harrisburg area nor Schuylkill County (plus the adjacent areas of Berks, which are demographically similar) has enough population for an entire district. This district is a Republican one, with Trump winning it 56-40 and McCain topping Obama 51-47. However, Democrat Tim Holden represented a district with similar partisanship during the 2000s. He’s probably the only Democrat who could win this district though.

District 12 is the Lehigh Valley district, containing the counties of Lehigh and Northampton and then, in an effort to avoid splitting any counties that aren’t already split, it picks up a slice of Berks. The district is still a swing district, as any Lehigh Valley-based district will be. Despite Obama’s 13-point win here in 2008, Trump won it by a single percentage point, and it has a very slight Republican lean (PVI of R+0.6). The Lehigh Valley is an obvious community of interest, and so a district similar to this will exist on pretty much every proposal. This quintessential swing district will undoubtedly be heavily contested by both sides.

PA 18 dis 2016 legal no gerry correct philly inset

District 13 contains the remainder of Berks County, including Reading and most of its suburbs, along with (almost) the entire remainder of Chester County. It’s the best way to deal with Chester and the Reading area so that Montgomery County has a district entirely within it and Delaware County is entirely contained in a district. This district has a slight Democratic lean (PVI: D+1.7) in presidential elections, but Republicans are still strong downballot. Obama won it 56-43 in 2008 and Clinton won it 52-44, buoyed by strong support in Chester County. Both parties would heavily contest this district.

District 14 is the district that contains all of Delaware County that every non-partisan redistricter agrees should exist. In the case of my map, it fills the remaining population by taking in the Merions in Montgomery County (which has to be split anyway), one ward of Philadelphia, and a few townships in Chester County. It’s solidly Democratic, having voted for Obama in 2008 by 27 points and Clinton by 29 points.

District 15 is the Bucks County district, plus two wards of Philadelphia. It is a very swingy district, having voted for Clinton by just one percent and having an R+0.6 PVI. It moves only a point or two to the left from the current version, which includes some Republican townships in upper Montgomery County – those are kept in the 16th in my proposal so that the 16th is contained entirely in Montgomery. District 15 would most likely be a battleground district in almost every election cycle.

District 16 now has a much cleaner shape. It’s entirely in Montgomery County, and includes the entire county minus the Merions. Despite dropping the Merions (which are heavily Democratic), the 16th still has a strong Democratic lean due to places like Cheltenham, Abington, and Norristown. It voted 59-40 for Obama in 2008 and 57-39 for Clinton, and has a PVI of D+6.2. Republicans would probably put up a strong candidate here (they do still have a substantial bench in Montgomery County), but Democrats would be strongly favored to win here.

District 17 is one of two districts entirely within Philadelphia, generally taking in the areas north of Center City and east of Broad Street. It is majority-white, but is over 25 percent African-American and also has large Hispanic and Asian populations. It is worth noting that not a single ward of Philadelphia is split in this map. This district would obviously elect a Democrat, and Republicans wouldn’t seriously contest it.

District 18, the final district on the map, is also entirely within Philadelphia, and has a population that is majority African-American. At 88-10 Clinton and D+38.9, it would be one of the most Democratic districts in the entire country. It consists of South Philly, West Philly, and part of northwest Philly. Its member of Congress is almost guaranteed to be an African-American Democrat.


Overall, this map has 11 Trump districts and only 7 Clinton districts, however two of those Trump districts voted for him only very narrowly and another one has a Democratic PVI, while one of the Clinton districts has a Republican PVI. The map creates five districts that are clearly competitive, with a couple others that could be competitive in the case of a massive wave or a particularly strong candidate. Considering that Pennsylvania narrowly voted for Trump and has voted for Democratic presidential candidate in the recent past, that breakdown is very well reflective of the state as a whole.

Due to the compactness of the districts, the fact that so few counties are split, and that it favors neither political party, the Pennsylvania Supreme Court should implement this map as the Congressional District map for Pennsylvania for the 2018 and 2020 elections.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s