How Much Will Redrawn Pa. Map Affect the Midterms?
The Supreme Court of Pennsylvania recently revealed its newly drawn maps for the commonwealth. If you’re interested in a down-in-the-weeds analysis of what the court’s 5-2 Democratic majority did with each congressional district, I recommend this explainer from the New York Times.
The bottom line is this: The court acted largely within the confines of its initial order, which demanded compactness, contiguity and minimal jurisdictional splits. Within those confines, however, it repeatedly made choices that increased the Democrats’ odds of winning districts.
For example, under the old map (and under maps extending back to 1942) Pennsylvania’s 8th Congressional District was largely based in Bucks County, a suburban/exurban county to the northeast of Philadelphia. To achieve population equality, that version of the district also took in parts of Montgomery County -- and those portions leaned Republican.
The court’s map, by contrast, places areas of Montgomery County that lean toward Democrats in the district, making it bluer. This doesn’t really serve any of the criteria the court announced, but it does make the district (now numbered as the 1st) about a point more Democratic. Indeed, had the court selected areas near the Montgomery/Bucks border somewhat to the north of the current lines or somewhat to the south, the district would be redder than it is under the current confines of the map.
Northeastern Philadelphia, an essentially suburban area that probably makes more sense appended to either the 8th or to the 13th District (which is based in Montgomery County, and is now renumbered as the 4th), is now placed into a district that approaches downtown Philly. This unusual choice set off a chain reaction to the southwest of the city: The old 7th District (now numbered the 5th) is based in Delaware County, but is pushed into southern Philly.
I didn’t think the court would make this move, because it in turn requires the 7th (5th) to be split among three counties, but I guess the court was okay with this. The court then selected parts of Montgomery County to place into the 7th (5th) rather than parts of Chester County. This then creates a three-way split of Montgomery County, and means that the Chester-based 6th District (still the 6th) takes in more of southeastern Chester County, which tends to lean Democratic.
The 6th is then pushed, somewhat incongruously, up into Berks County (to the north), when it could have been made more compact – and more Republican — by expanding into rural Lancaster County. Berks County has plenty of “red” in it, but the district’s appendage only goes up into the heavily Democratic working-class town of Reading. The district now has a Democratic lean.
We can go on: Democrat Matt Cartwright’s district could have been almost perfectly drawn between Lackawanna, Luzerne, and Monroe counties, which would have pushed the old 15th (now numbered the 7th) into heavily Republican Carbon County, instead of into Democratic-leaning areas of Monroe County. In other words, overall, this is probably about as good a map as Democrats could hope for (indeed, some analysis suggests that this is an even better map for them than the one that the Democratic legislature proposed).
Does this then count as a Democratic gerrymander? It’s an interesting philosophical question and one that lies at the heart of the current debate over gerrymandering. Imagine that in a roughly 50-50 state, we were able to see all of the possible compact maps that didn’t split political boundaries. Suppose that the average map would elect 11 Republicans and seven Democrats in a neutral year, that the most Republican-friendly map would elect 12 Republicans and six Democrats, while the most Democratic-friendly map would elect nine Republicans and nine Democrats at the other extreme.
Given this, is a map that will tend to elect 10 Republicans and eight Democrats over time a gerrymander? In one sense, it is a Republican gerrymander of one seat, because we would theoretically like to see nine and nine elected in a 50-50 state, and Republicans fare one seat better, notwithstanding the fact that the nine-nine map is possible. In another sense, it is a Democratic gerrymander of a seat, because a process that removed politics from consideration would tend to see Republicans do a bit better than the given map will.
There’s no real answer to this question. It is a normative question, and one where it is probably impossible to extricate oneself from one’s partisan preferences. Unsurprisingly, liberals tend to emphasize the fact that the map pulls the state closer to a nine-nine equilibrium, while conservatives emphasize the individual choices that the mapmakers made. Of course, we also cannot set aside the fact that this map was drawn by, in effect, elected Democratic politicians, but that is a broader problem with elected judiciaries in general.
But how much will this matter in 2018? Probably less than people anticipate. Remember, Republicans were very likely to lose seats in Pennsylvania regardless of how the map was redrawn. In a bad Republican year, they probably would have lost multiple seats.
To give a sense of how this would play out, consider the following thought experiment. Assume a neutral year, where Democrats basically win all of the districts that lean Democratic, while Republicans win all of the districts that lean Republican. Let’s also assume that Democrat Cartwright would win his nominally Republican district regardless of the redraw (these sorts of anomalous results represent challenges for any measurement of gerrymandering based on electoral outcomes, but that is another story).
Under the current map, there are four districts that lean toward Democrats, plus Cartwright’s. Under the redrawn map, there are seven districts that lean toward Democrats, plus Cartwright’s. Thus, in an even year, Democrats would gain three seats under the new map that they would not have otherwise won.
We can then push this up in increments, to see if and how these numbers change at different levels of Democratic support. So, for example, if Democrats were to win seats that leaned Republican by 0.5 points or less, they would still pick up no seats under the current map, and three under the redrawn map. I’ve given the results for waves of different sizes below (results under old and redrawn map are taken from Dave Wasserman of Cook Political Report):
As you can see, in neutral years the redrawn map results in a nice advantage for Democrats over the current map. Rather than the current 14-4 Republican map (plus one for Cartwright), they would have an 11-7 Republican map (plus one for Cartwright). In a year where Democrats were able to swing all of the seats that lean Republican by a point or less into the Democrats’ camp, they would probably pick up four seats (and hold Cartwright’s).
But as the wave grows for Democrats, the advantage for the current map shrinks. In an election where Democrats win all of the districts that lean Republican by three points or more – which is roughly the environment they need to win the House, this map only nets them one seat over what they would expect to win under the current map.
It is also worth noting that in a reasonably good Republican year – one like 2010 or 2014, where Republicans perform well in seats that are two points more Democratic than the country as a whole – we would expect a 13-5 Republican map. In other words, we’d be right back where we are today.
So, in the big picture, the effect of this redraw is overstated in terms of the battle for the House. In an environment where control of the House is likely to be a question – and to be clear, we’re absolutely in such an environment today – Pennsylvania Republicans were likely to lose many of these House seats in the first place.