In past competitions, the task has been to draw a congressional map, with seventeen or eighteen districts. Here, that is completely changed, in the quest to draw a map for the state House of Representatives. While I drew the congressional districts in the name of competition above all else, that goal is somewhat changed here.
In congressional races, a ten-point separation – or a PVI of five points – is about as far apart as a district can be and stay competitive. This is because of how partisan it has gotten. The attitude is less “I’m voting for [insert representative], and more “I’m voting Democrat/Republican.” When the debate is so polarized, in order to ensure that every voter has a say – and to make a district that could truly go either way – a close district is needed.
But in the state house races, the districts are smaller, a lot smaller, and there’s a lot more personal contact. Representatives are more likely to know their constituents personally in districts of 60,000 than in districts of 750,000. And because of this, although many people keep the same attitude – “I’m voting for [insert party]” – lots of people also feel a connection to their local representative and will vote accordingly based on that.
It’s also a lot harder to draw ultra-competitive districts – those with a PVI of less than five points either way – with smaller districts, while still keeping similar-minded communities together and maintaining compact districts. So keeping this in mind, I set a bar for myself – while optimally the PVI would be under five, a PVI under ten would still be acceptable given the greater elasticity of smaller districts.
There are situations in which we can see this to be true. In my map’s 5th district – in northwest PA, serving the western bit of the Erie area, Meadville, and western Erie and Crawford counties – the PVI is R+7.18, outside the ideal range for competitive districts. Yet in 2008, in the presidential race, John McCain carried the district for the Republicans by only 264 votes, or a margin of 0.9%. With a popular local Democrat, this would be a very winnable district. Similarly, in the southeast of the state, the 182nd district has a PVI of D+5.91, but the area – encompassing Whitemarsh, Ambler, and other areas of suburban Montgomery County – is fertile ground for moderate Republicans, who have a long history of winning in the area.
Sometimes, in all honestly, towns and cities did have to be split to maintain competitive districts. One of the best cases of this was in my town, State College, and the surrounding area. This area could either have been put in two safe districts – one for each party – and a third R+8-ish seat. However, all three could be made competitive by splitting State College into three, pairing each part with rural areas outside of town, and creating three competitive districts – two pure tossups and one leaning Republican. I chose the latter – I decided that in my view, it was more important for every voter to have a real say in the election than to have a single representative serving a given town or city. I applied this logic elsewhere too, but still tried to keep built-up areas within a single district. Competitiveness, though, was the most important principle. And in that regard, if forced to choose between splitting a town and creating a safe district, I chose the former – every single time.
Ultimately, although the “analysis” tab in Dave’s Redistricting provides us with a count of the number of competitive districts, I didn’t use it. Instead, I used the PVI which was computed for each district. This is a better estimate of the actual expected result, as it takes into consideration the 2012-16 election results instead of just the 2008 presidential results – what DRA uses – which produced a landslide for Obama and throws the numbers out of whack. Using 2012-16 PVI allows us to see the voting trends in an average year, which is what we want to do to create competitive districts.
The final count I reached was 113 competitive districts against 90 safe ones, and of the 113, 64 were ultra-competitive – the PVI was under five points. Given the difficulty of drawing competitive districts, 113 was as many as I could come up with without seriously distorting the look of the map.
Overall, drawing this map was a fascinating exercise. While the congressional maps were purely logistical puzzles – “how many tossups can I cram in?” – this one was not. This map involved a lot of thought about what values ought to be prioritized, and in which order, in such a way so as to maximize the democratic input of the people on their legislature.
Drawing a state legislative map also convinced me further that the current electoral system is woefully inadequate. Our first-past-the-post system where whoever wins the most votes wins the seat produces safe seats where voters don’t matter and creates a large possibility that a party losing the popular vote can win a majority in the legislature. Most developed countries worldwide use a system of proportional representation, where seats are allocated according to the proportion of the vote earned by each party, meaning that “safe seats” are a thing of the past and that every voter’s voice counts equally. The fact that almost half of the state would still, despite my best efforts, have no say in choosing their safe-seat representative shows that our system is no longer fit for purpose.