One of the biggest scandals in the US is how districts are drawn to benefit the party in power, to give them legislative representation that is much greater than their share of the vote merits. This is because the party that controls the legislature at the time of the drawing every ten years pretty much gets to draw the lines as it wishes as long as the area is contiguous and contains roughly equal numbers of voters. They thus can draw the most outlandishly shaped districts in order to concentrate most of their opponents’ votes into a few districts. Only the grossest abuses, the ones that use race as the main factor, get rejected by the courts, because race gets the highest standard of scrutiny. But if race is not an obvious factor, you can with impunity shepherd the supporters of the opposing party into a few districts.
Finding a way to draw districts fairly has been challenge but this method seems promising because it is simple and easy to understand.
With the new approach, one political party gets to draw an electoral map that divides the state into the agreed number of districts. The second political party then chooses one district to freeze so that no more changes can be made to it by either side. They then get to redraw the rest of the map.
Once the new map is complete, the first political party freezes one of the new districts so that no further changes can be made to it, and is allowed to redraw the rest of the map again. This process goes back and forth until every district within the state is frozen. In Pennsylvania, for example, this would require 17 cycles as there are 18 districts.
But the real problem does not lie in finding a fair way to do this. There are many such suggestions and this one seems to be one of the better ones. The problem is that the party in power is usually not interested in fairness but in increasing their power, so getting such a system adopted by the legislature will not be easy. This may be one of those rare cases where a constitutional amendment that is voted on by the people directly is the most desirable option.
A nice variation on the “whoever cuts the pie gets the last slice” formula, but one that practically guarantees arbitrary division of local communities.
Surely software algorithms could produce multiple suggested divisions accommodating the need for compactness, simplicity, transportation availability, relevant court mandates, etc, and ad-hoc multi-(or non-)partisan commissions could then select among them.
Something like that (sans map-savvy computers) was the process here in Florida, until the Democrats then controlling the state legislature decided to arrogate that function for themselves. Then the Republicans’ “southern strategy” swept them away, and the D party has held the minority position ever since.
How about setting a geometrical limit on the shape of the district? A pure number that is totally independent of politics? (OK, that may be a problem…)
Say, define a Gerrymandering Index as the length of the border squared, and divided by the area. For a square the GI is 16, for a 1:2 rectangular box 18, etc. There probably aren’t any good reasons for shapes whose GI is above 50. Except maybe some weird natural borders like meandering rivers.
The proffered solution contains the assumption that there are two major parties, who get to divvy up the spoils. The right to fair representation belongs to the people, not to political parties.
There is growing interest in attacking the problem mathematically.
There are various algorithms for splitting up territory into districts. Most involve Monte Carlo methods, sometimes enhanced by genetic algorithms.
Quantifying Gerrymandering
The next step is to score the solution for fairness:
Here’s How We Can End Gerrymandering Once and for All
When such methods are used to score current districting lines, numerous states are found to be unfairly biased in favor of one party:
How the New Math of Gerrymandering Works
A quote from that last article:
In (AFAIK) every other country with a non-ridiculous claim to hold democratic elections, redistricting is the responsibility of a non-partisan, appointed body, composed of relevant experts. Obviously this isn’t foolproof, but it seems to work a lot better than giving the current ruling party carte blanche; and the proposed system would make it even more difficult for the party duopoly in the USA to be challenged. (If that’s possible -- the dice are already weighted so heavily against such a challenge that the current incumbent parties have maintained their absolute dominance for a century and a half.)
I have thought about use of geologic boundaries as political boundaried, because to a crude approximation, people living on the same formation would have have common land use goals and resources. The devil would be in the details though and formation contacts can be very complicated. And on the flip side it would politicize geologic maps and politicians would try to override sound geologic decisions to their benefit. Sound familiar?