Alan Zendell, May 29, 2017

National polls have long shown that voters decry Congressional gridlock and ultra-partisanship. Arguably, they were a major part of the general displeasure with Washington which fueled the populist uprising that drove the campaigns of Donald Trump and Bernie Sanders. Candidates talked about forcing Congress to reach across the aisle to get things done, but that was just a pipe dream. None of the campaign bombast offered any suggestions about how to fix the problem. That’s because the only way to fix it is to place limits on gerrymandering. It would be nice to eliminate it entirely, but that’s an even more unrealistic fantasy.

The Supreme Court has already acknowledged that gerrymandering is a serious problem. Recently, they overruled the way North Carolina drew its Congressional boundaries, finding that they discriminated based on race. That’s an important precedent, but the larger problem in terms of overall numbers is arranging districts lines to favor one political party over another.

The Court will be looking at this issue in the near future. They have already stated that when it can be shown that drawing Congressional lines based on partisanship is extreme, it is unconstitutional, but no one has been able to propose a quantitative method for determining what crosses the line.

It’s one of a class of interesting problems that are always thorns in the sides of mathematicians and statisticians. Anyone can see just by looking at a map that the lines are unreasonable. Some districts are so skewed and misshapen they are a constant source of jokes, but intuitively knowing that something is wrong doesn’t mean the problem has an obvious solution, as any student of geometry can attest. Statistics is rife with problems whose solutions defy intuition.

Most observers accept that gerrymandering will never be entirely eliminated, but it may be possible to develop standards and methods that determine when excesses become illegal. A possible solution is discussed in https://newrepublic.com/article/118534/gerrymandering-efficiency-gap-better-way-measure-gerrymandering, which addresses the concept of partisan symmetry that was raised by Justice Stevens in a 2006 case. As co-author Nicholas Stephanopoulos explained, partisan symmetry requires that partisan votes translate into seats with equal efficiency, and several justices suggested that this approach might find a favorable reaction from the Court in the future.

The article referenced above discusses such an approach based on measuring a concept called the “efficiency gap”. It’s an interesting notion that counts wasted votes; that is, votes in excess of what the winning candidate needed for victory as well as every vote cast for the loser. The idea is that gerrymandering attempts to minimize the wasted votes of the winner while maximizing those of the loser. The question is whether a test can be developed to define when that practice is so extreme, it becomes unconstitutional.

The efficiency gap measures the difference between votes wasted because of packing (creating unnecessarily large majorities) and cracking (causing too many votes to be wasted by losers). The ideal would be “a state with perfect partisan symmetry, [in which] both parties would have the same number of wasted votes.” Stephanopoulos says the efficiency gap represents the winning party’s “undeserved seat share.” It sounds simple and fair, but that doesn’t mean it’s easy.

The authors studied elections from 1972 to 2012, finding that the efficiency gap peaked in 2012, and that in seven large states it resulted in an unfair swing of more than two legislative seats, and every one of those seats favored the Republican Party. They went on to propose (somewhat arbitrarily) that a reasonable standard for determining when gerrymandering is extreme enough to be unconstitutional is when it produces two or more undeserved seats, and further, that the system in place is “unlikely to fade away” given plausible assumptions concerning future vote swings. Their hope is that lower courts would find this test consistent with the comments made by the Justices in 2006, which would make it easier for the high court to agree.

Could this happen in any of the upcoming cases on the Court’s agenda? It’s impossible to predict. There is always a debate between those who favor a strict interpretation of the Constitution and rulings that appear to adapt the words of the Constitution to address issues not foreseen by the founders. Given the role gerrymandering has played in causing our government to become dysfunctional, and the likely importance of next year’s congressional election, the debate ought to be interesting.