Pennsylvania sends too many Republicans to Washington.
That's not a partisan attack. It's just math.
Of the 18 Pennsylvanian members of the House of Representatives, 13 are Republicans and 5 are Democrats. That split should be more like 11 to 7 or even 10 to 8 if the districts were drawn without attempts at favoring Republicans, according to recent expert analyses.
It's all about the map: Several lawsuits are attempting to get various state legislative and congressional maps declared unconstitutional on the basis of partisan gerrymandering, the idea that one political party drew the lines in a way that benefited them unfairly.
The lawsuits rely on a set of tools that for the first time could convincingly identify skewed maps and persuade the courts that a state's map goes too far in favoring one party. A federal court has ruled Wisconsin's state legislative map unconstitutional, the first victory in a partisan gerrymandering case in three decades.
That decision used one of several new mathematical tests to help measure the map's Republican skew, and the Supreme Court will hear the case in the fall; if it upholds the decision, it could create a legal standard, potentially including some of these tests for measuring map bias.
That could spell trouble for Pennsylvania, which fails those analyses.
People have made these claims before, but proof has been elusive. The Supreme Court had said too much gerrymandering could be unconstitutional, but the justices couldn't agree on how much is too much — in part for lack of measurement standards.
Anthony Kennedy, the pivotal swing vote in the 2004 Pennsylvania case, laid down a gauntlet: A convincing test hadn't been found, but that didn't make it an impossibility. That opinion — and how close the court came to declaring partisan gerrymandering a political issue off-limits to the courts — spurred academics and lawyers to put forward a host of mathematical methods to precisely identify skewed maps.
"No matter what concept you care about in partisan gerrymandering, Pennsylvania is going to be an outlier," said Eric McGhee, a research fellow at the Public Policy Institute of California who helped develop the "efficiency gap" test.
Unlike in many other states, Pennsylvania's congressional district map is drawn by the state legislature, passed as a bill, and signed by the governor. The current map was drawn in 2011 by Republicans, who controlled both houses of the legislature and the governor's mansion. New Jersey, which uses a commission of political appointees, has a bipartisan split and tiebreaker vote; its map is generally not flagged as a problem by these tests.
The 435 seats in the House of Representatives are divvied up after the census every 10 years, based on population. As populations shift across the country, so, too, does political power — Pennsylvania has lost at least one seat every 10 years, while states in the Southwest have grown. Once the state is given its number of representatives, it redraws its map.
Those mapping decisions ultimately can shape government policies that affect millions of Americans.
"Pennsylvania is clearly quite extreme. … This is not random," said Michael Li, a redistricting and voting rights expert who is senior counsel for the Democracy Program at the Brennan Center for Justice at New York University School of Law.
According to tests run by Li and others, the 2011 redrawing of Pennsylvania's congressional districts skewed the districts in Republicans' favor, giving two or even three seats to Republicans over what would occur with a politically neutral map.
Here is a summary of how Pennsylvania fares in three of the tests that experts believe best capture evidence of gerrymandering.
The primary gerrymandering methods are "packing" and "cracking."
In packing, a rival party's voters are concentrated in a district it usually wins easily. While that district's race is conceded, the rival's voters can't cast ballots in elections that would be more contested. Cracking involves dispersing a party's voters into multiple districts in such ways that they are deprived of majorities.
By their measures, the Keystone State should be sending 10 Republicans and 8 Democrats to the House — not 13 and 5.
Princeton University researchers use "very well-known and very well-tested, battle-tested statistical tests" to measure ways that mapping can affect elections, said Brian Remlinger, the statistical research assistant who serves as the main analyst at the Princeton Gerrymandering Project.
Cracking and packing would result in one party's winning districts with wide margins, while the other party wins more districts, but with slimmer margins. That effect was evident in Pennsylvania in 2016, when winning Democrats on average took 75 percent of the vote, compared with 64 for Republicans. The odds of that happening would have been less than three percent with a politically neutral map, Remlinger said.
Historic seats-to-votes curve
Historically, what percentage of votes did a party receive in Pennsylvania, how many seats did it win, and how does that compare to recent elections?
Pennsylvania's map is the most heavily skewed in the country on this measure, a Brennan Center analysis found. In 2016, Republicans had about a four-seat advantage, compared with the outcomes in a neutral mapping, and about five seats in 2012.
Implications for Pennsylvania
"There isn't necessarily a need to pick one test, you can have multiple tests and the fact that multiple tests point in the same direction, as in Pennsylvania, suggest this isn't random," Li said.
The Supreme Court will hear arguments this fall in Gill v. Whitford, the Wisconsin case in which a three-judge panel relied on the efficiency gap, in part, to declare Wisconsin's state legislative map unconstitutional.
If the Supreme Court upholds that decision, using the efficiency gap or other measures, that in effect would green-light the use of these measures in gerrymandering case law, experts said, and open the door for other maps to get challenged using these tests.
"All of these things tend to point in the same direction, so we think that presenting any of these pieces of evidence could be useful for a court," said Ruth Greenwood, senior legal counsel at the Campaign Legal Center, which is bringing the Whitford case before the Supreme Court.
Pennsylvania's map is already facing a direct challenge: The League of Women Voters of Pennsylvania in June sued the state, using the efficiency gap as part of its legal argument that partisan gerrymandering had occurred.
"An election system in which one party, whatever party happens to be in control of the election system, can rig it so it can win and keep a majority thereafter, is almost by definition the antithesis of self-government," said Michael Churchill, an attorney at Philadelphia-based Public Interest Law Center, which is representing the League in the case.
"It defeats the very purpose of having elections."
Correction: A previous version of this story misstated the probability of Pennsylvania's results occurring with a politically neutral map according to the lopsided wins test because of erroneous data provided by Princeton Gerrymandering Project's web site. The odds are less than three percent.
Gerrymandering is an age-old process of dividing congressional and legislative districts in such a way as to give one party an advantage.
On Feb. 11, 1812, Gov. Elbridge Gerry, then the governor of Massachusetts, signed into law a redistricting plan aimed at keep his party in power.
The Boston Gazette printed mock map in the shape of a salamander under the headline "The Gerry-mander."