Guest Column by SUNY Binghamton Professor Jonathan Krasno
NYS’s congressional map has been derided as a partisan gerrymander by both Republicans and nonpartisan observers, mostly because Democrats are poised to gain seats under the new lines. Changes from the status quo is a poor way to evaluate any map. The facts here suggest that NYS’s map is not a partisan gerrymander by Democrats.
The problem with using the status quo as a baseline is obvious: it might be biased, too, no matter who created it. In the case of NYS’s congressional map, the mapmaker was officially nonpartisan, a federal magistrate judge. Lack of partisan motives is no guarantee of nonpartisan outcomes particularly in a state like NY.
The issue in NYS is the extreme concentration of its residents and Democrats in NYC. NYC makes up 44% of the state population (its metro area comprises 68%) and is overwhelmingly Democratic. NYC is unique, too, in that it is squeezed like the narrow part of an hourglass between the expansive territory of northern part of the state where 42% of the population resides and Long Island to the south where the remaining 14% lives.
That geography makes it difficult to draw a map that spreads many of NYC’s residents to other districts outside the city. That might seem like an odd thing to want to do, but it’s the core issue in creating a congressional map that reflects the political complexion of the whole state rather than region by region.
Here’s a simple illustration why using the current map. Trump’s share of the vote in the 8 House districts currently held by Republicans was @55% in 2016, @57% in 2020 – comfortable but hardly overwhelming margins. In the 19 districts held by Democrats, however, Clinton and Biden won @69% and @72% respectively. Blowouts.
That disparity reflects what in gerrymandering is called “packing,” putting a large number of one side’s voters in a small number of districts while spreading out the other side’s in a larger number. It turns out that NYS is so strongly Democratic that the congressional map packs Democratic voters and still leaves Democrats with 2/3rds of seats. But it’s also true that Democratic voters get a raw deal here in that their votes essentially count less than do Republican ones when it comes to winning House seats.
Does the new map do better? NYS loses one of its 27 seats, dropping the number to 26 and the expected partisan split becomes 22 to 4. In the 4 Republican districts Trump’s vote share increases to @64% in 2016 and @59% in 2016, while in the 22 Democratic districts Clinton’s vote remains @69 in 2016 and Biden’s falls a few points to @68% in 2020. The packing still favors Republicans, though by a little less.
That impression is supported by closer analysis. The easiest way to assess packing is to compare a party’s vote share in the median district to its average vote across all districts. The median is important because it reflects the pivot point in any legislature, the seat that must be carried for a party to win a majority of seats. The reason why a median and mean might be different is because one side has had a lot of its voters packed into a small number of districts.
This insight isn’t particularly profound; the idea first originated with a 19th century statistician who suggested it as an example to explain skewed distributions. For non-mathematical purposes this approach has the advantage of being especially easy to calculate and interpret.
In NYS’s proposed House map, Democrats’ presidential vote in the median House district was 57.5% in 2016 and 60.2% in 2020. If the map was fair, their mean vote across all 26 districts would be very close to the median. If the map was biased in favor of the Democrats, they would do better in the median district because they would have gerrymandered Republicans into a small number of districts.
In fact, the third option is what occurs – Democrats do better across the 26 districts on average than they do in the median district. The disparity is fairly large, too. Democrats averaged 64.1% across the 26 districts in 2016 and 63.8% in 2020. The difference between median and mean represents the anti-majoritarian bias in the map. For instance, in 2020 the difference between Democrats’ share in median and mean – 60.2 vs 63.3 = 3.1 points – reflects how much the party needs to win in excess of 50% to likely carry the median.
The actual numbers here take some time to digest, but the bottom line is simple. Just because NYS’s new congressional map has Democrats picking up seats doesn’t make it a partisan gerrymander. Closer analysis suggests that, if anything, the new map does only marginally better at resolving the geographical challenge posed by the concentration and location of NYC with its millions of voters. The new map is no Democratic gerrymander.