Last month the Pennsylvania Supreme Court instructed the state's Republican-led legislature to draw a new congressional map after finding the existing one was an illegal partisan gerrymander that violated voters' right to participate in “free and equal elections.”
On Friday, Republican leaders in the legislature submitted their new map for the governor's approval. As directed by the Supreme Court, the new map is much more compact than the old one. Gone are the infamous convolutions that characterized the old map, earning nicknames such as “Goofy kicking Donald Duck.”
The new districts generally respect county and municipal boundaries and don't “wander seemingly arbitrarily across Pennsylvania,” as the state's Supreme Court wrote. Unfortunately for Pennsylvania voters, the new districts show just as much partisan bias as the old ones.
You can demonstrate this using the precinct-level results of the 2016 presidential election: See which precincts are assigned to which districts under the new map, use those assignments to calculate the total presidential vote in each of the new districts and compare those figures with the vote totals under the old districts. That will give you a good sense of how the partisan makeup of the new districts compares to the old ones.
Brian Amos, a redistricting expert at the University of Florida, has done exactly that. Amos combined the new district maps with precinct-level returns compiled by cartographers Nathaniel Kelso and Michal Migurski.
The similarities are striking: In 2016, Donald Trump received more votes than Hillary Clinton in 12 out of Pennsylvania's 18 districts. Under the Republicans' new map, Trump would similarly outperform Clinton in exactly 12 districts.
Not only that, but the vote margins in each district would be virtually identical. The chart below plots, for each district, the vote margins in 2016 vs. the margins that would result from the Republicans' new map. Across all 18 districts, the average difference in vote margins between the old and new map would be a little over four percentage points.
From a partisan standpoint, in other words, the new map is almost exactly like the old one. Under the existing map, Democratic House candidates have routinely received roughly 50 percent of the statewide popular House vote but only five of the state's 18 House seats. The new map is unlikely to change that.
“This reminds me of what happened in Florida in 2012, where the state legislature drew maps under new constitutional anti-gerrymandering requirements,” Amos said. He should know: Amos was a consulting expert for the plaintiffs in the lawsuit that ultimately overturned some of Florida's gerrymanders.
In that case the legislature redrew maps that were more compact but just as biased. “Compared to the previous decade's plan, it was an improvement in measures like compactness and respecting county and municipal boundaries,” Amos said. “But there was still a strong Republican bias, which is why the congressional and State Senate plans were struck down for being gerrymanders.”
Pennsylvania Republicans deny partisanship was a factor in their new maps. “Partisan data was not used in any form in the development of the proposed map,” Drew Crompton, general counsel to Pennsylvania's state Senate Republicans, said in an email. “The proposed map fully complies with all the metrics laid out by the plaintiff experts in the case as well as the tests set forth in the majority opinion by the Supreme Court of Pa.”
The plaintiffs aren't buying it. “Our experts have analyzed this map, and under the criteria set by the Supreme Court and common sense, their proposed map fails,” said Mimi McKenzie, legal director of the Public Interest Law Center, which is representing the League of Women Voters in its lawsuit against Pennsylvania's maps. The new map “exhibits extreme pro-Republican bias that mathematically can only have happened through an intentional effort.”
Amos, who is not involved in the Pennsylvania lawsuit, pointed out that you don't need partisan data to produce a partisan map. “If you have a strong interest in your state's politics, you've probably looked at detailed maps of election results before, if not studied them closely,” he said. “It might not be as precise as getting exact party breakdown predictions, as is possible with redistricting software loaded up with those data, but it doesn't guarantee that the process was partisan-blind.”
Experts consulted by the plaintiffs in the Pennsylvania case have testified that it's highly unlikely the level of partisan skew observed in the old maps — and now the new ones — is accidental. Jowei Chen of the University of Michigan created 500 computer-drawn district maps optimized for Pennsylvania’s traditional redistricting criteria of compactness, contiguity and equal population. Among those maps, “the highest number of classified Republican districts was 10, and in none of the simulated plans would 13 of the congressional districts be classified as Republican” as they are under the existing maps, Chen found.
Pennsylvania Gov. Tom Wolf (D) has until Thursday to review the new Republican-drawn map. If Wolf rejects it, the Supreme Court will instruct independent redistricting expert Nathaniel Persily of Stanford University to draw a new map from scratch.