Do Democrats Stand to Gain From Redrawn Pa. Maps?
On Monday, the Supreme Court of Pennsylvania struck down the state’s Republican-drawn congressional map. Because it did so on state constitutional grounds, there is likely no appeal to the U.S. Supreme Court. Democrats immediately celebrated. Marc Elias, Hillary Clinton’s former counsel, tweeted: “Pennsylvania Supreme Court strikes down congressional map. Orders new map in time for 2018 election. Since the case was decided under the State constitution, there is no route to SCOTUS. There will be a new, fair map for 2018.”
I’m not sure this is correct. It’s hard to say without seeing the full opinion — and that is a very important caveat — but for right now, it appears that the court didn’t order fair districts so much as it ordered compact ones. This is an important distinction.
Perhaps because the 2018 elections are fast approaching, the court did not publish a full opinion, but rather entered an order, with the opinion to follow. After mandating that new maps be drawn by Feb. 9, the justices presented the following limitations on those maps: “any congressional districting plan shall consist of: congressional districts composed of compact and contiguous territory; as nearly equal in population as practicable; and which do not divide any county, city, incorporated town, borough, township, or ward, except where necessary to ensure equality of population.” Now, it’s entirely possible that the eventual opinion could insist that one of the plaintiff’s proposed fairness metrics, such as the efficiency gap, be incorporated into a new map, which would force something like a map electing nine Republicans and nine Democrats, regardless of geographical constraints. But it would be strange to set a three-week deadline for the legislature, and then add additional constraints halfway through.
So for now, at least, a map has to be compact, equi-populous, and most importantly, must not divide political subdivisions except when needed to ensure equality of population. If the court really means what it says with this last requirement, then there are probably only a handful of maps that can actually be drawn. Moreover, the map that comes out of this may end up strengthening the GOP’s position in an environment such as this one.
To see what I mean, I will focus in on Southeast Pennsylvania, where most of the competitive districts in the state reside. I’ll pay the most attention to Philadelphia, Delaware, Chester, Montgomery, Bucks, Northampton, Lehigh, Berks and Lancaster counties, which you can see in this map of southeastern Pennsylvania.
Here is what the current districts in this area look like. They are quite a mess. Indeed, it isn’t difficult to see why the court wanted to strike them down, even if we don’t know exactly why it did so:
Using a tool known as Dave’s Redistricting App, which allows you to draw congressional districts, I’ll generate maps and show some of the limitations a strict interpretation of the court’s order imposes upon redistricters (again, with the caveat that we don’t know exactly what the criteria will eventually be). First, though, let’s detour to the northeast. Interestingly, there are three counties there – Luzerne, Lackawanna and Monroe – whose combined population is almost exactly the ideal size for a district. To comply with the order, you probably have to draw something like this district:
This is Matt Cartwright’s district, which was initially drawn as a Democratic “vote sink” – a collection of packed Democratic voters – but which has swung to the right over the past decade. This new district would actually lean Republican by two points. In this environment, it doesn’t pose a problem for him. If 2020 is a good Republican year, however, he could have problems.
We will begin by drawing Philadelphia districts. The city will hold two districts and part of a third, so we must draw at least two districts. The question is where the “leftovers” go. We could try putting the leftovers in the southwestern portion of the city, and combining them with Pat Meehan’s 7th District – the Delaware County-based district.
The problem with this is that the Delaware County-South Philly district is 30,000 residents shy of having enough people to create a full district. To make this work, you have to draw the 7th across three counties.
An interesting possibility would be to draw the two Philly districts like this, and leave the Mt. Airy/Chestnut Hill/Roxborough precincts for the 13th (the “bump” in the northwest). This probably technically satisfies the court’s compactness requirements the best, since the 13th basically becomes a box in this scenario; this also makes the 1st District minority-majority, which may be required by the Voting Rights Act.
But to keep things interesting, I’ll maintain the historical split, which puts northeastern Philly into its own area. Indeed, from 1902 through 2002, this traditionally Catholic, working-class area of Philadelphia had its own congressional district. From a “communities of interest” perspective (putting demographically similar areas with similar interests into the same district) this makes the most sense, although respecting communities of interest isn’t a factor to consider under the court order.
From here, there’s something of a fork. You can put NE Philly in the Bucks County-based 8th, or in the Montgomery County-based 13th. There are good reasons for and against this. If you put these voters in the 8th, you have to split Bucks County, which hasn’t been done since the 1870s. That said, this configuration is probably more compact than putting these voters in the 13th (although again, putting NW Philly in the 13th is probably what you do if you’re concerned about compactness). Regardless, I will play it out both ways. First, I will put these voters into the 8th. After you fill in southern Bucks County, the 8th moves a little less than a point toward Democrats.
Next, we have to decide what do to with the 7th. It becomes much more Democratic after shedding its Chester County exurbs and picking up precincts that had previously been placed in the 1st – it’s basically a guaranteed Democratic pickup. But Delaware County can’t support a single district on its own, so the 7th either has to extend into Montgomery County or into Chester County. If we put it into Montgomery County, the 13th has to go out into Berks County, and actually becomes somewhat marginal (about three points more Democratic than the country). The 15th district gets consolidated around Bethlehem and Allentown, but gets pushed into the redder areas of Bucks County and (probably) Carbon County. The 6th goes into Lancaster County. The results of this is a map that looks like this:
After filling in Pittsburgh, the rest of the state leans Republican by five points, and a lot of the remaining blue areas will go to finishing out the Pittsburgh seat. That’s not to say you can’t draw another blue district. It’s just hard, especially under the compactness criteria.
Alternatively, if we push the 7th into Chester County, the 6th gets pushed deeply into heavily Republican Lancaster County. The 13th pulls back into Montgomery County, and a new, swing-y 16th District probably has to be built around Reading. In this environment, the Democrats would probably win it.
Stepping back, if we had put northeast Philly into the 13th at step one, and pushed the 7th into Montgomery County, we get a map that looks something like this:
Note how the 8th District gets pushed toward Allentown, and the 15th now gets pushed into Berks. The district that contains Reading could be made into a swing district, but there are a lot of contingencies at work there. Most importantly, the three-way split of Montgomery County probably means this map isn’t acceptable.
Finally, if we had pushed the 7th into Chester County and put northeastern Philly into the 13th, we would get something along these lines:
What does this mean?
I’ve summarized the different partisanship orientations of the districts as follows. As you can see, there are tradeoffs under each map. The 7th always becomes unwinnable for Republicans, the 8th becomes more marginal, and the 6th and 15th tend to become harder to win for Democrats:
Here’s the real issue though: In a good environment for Democrats, a lot of these seats were set to flip anyway. So while this locks in a flip for Democrats in the 7th District, it also limits some of their upside. You can see this in the following chart. The way to read it is basically: “If the Democratic tide rises to ‘x’ level, how many seats would Democrats win in Pennsylvania?”
So, in a neutral year, the Democrats wouldn’t get any seats under the current plan (setting aside Meehan’s sex scandal), and would win one under the new plan (assuming Cartwright -- pictured -- holds on). If the environment were good enough to win R+1 districts, the first map that I drew would gain Democrats two seats (the 7th and the 8th), which is one more than other versions. But as you can see, the differences tend not to be substantial, and at higher levels of “wave,” most of the redrawn maps trail the current map for Democratic pickups. That’s because the Democratic vote in the state is so heavily clustered in Delaware, Montgomery, and Philadelphia counties (almost half of Hillary Clinton’s votes came from these three, plus Allegheny-Pittsburgh), and since you can’t unpack those counties under the court’s order, there are limits to how much Democrats are likely to benefit from this order.
Of course, there are caveats. A lot of this is calculated off of PVI, which includes Donald Trump’s strong performance in the state in 2016. If we assume this was a fluke, and not the culmination of a realignment of rural/exurban Pennsylvania, Cartwright’s district looks a whole lot more secure, while the 6th and 15th are a whole lot more “gettable” in a Democratic wave. But also recall that PVI is calculated across two elections, so Trump’s coalition is already somewhat discounted here.
Second, it bears repeating that we don’t have the court’s final opinion. It may incorporate some “fairness” considerations, or loosen the compactness/splits requirement.
Third, and most importantly, there is a Democratic governor in the state. He will likely push for “fair” districts. But again, the court’s compactness and line-splitting requirements limit somewhat the ability to make this happen. If the parties can’t come to an agreement, the court will draw the lines. It’s tempting to say that the elected 5-2 Democratic majority on the court would select/draw a highly favorable map for Democrats. But they just passed on an opportunity to read something like the efficiency gap into the state’s constitution, which would have effectively mandated that Delaware/Montgomery/Philadelphia always be unpacked to achieve partisan symmetry. They also appear to have foreclosed the Democrats’ ability to string out districts from the inner suburbs to the exurbs (“baconmandering”) when they eventually control the process (possibly in 2020). I’d initially assumed the court would select a highly favorable map for Democrats. After contemplating the order, I’m not sure.
Perhaps most importantly, all of this is quickly drawn, with just a day or so to think about it. Who knows what mapmakers will come up with in a week or two dedicated to nothing but this question. My goal here isn’t to be exhaustive. Rather, it is just to emphasize that the court didn’t mandate fair districts, but rather ordered compact ones, and that these are two very different things. Moreover, there are important implications to this distinction, and they don’t necessarily work in Democrats’ favor in every environment.