It is possible to end Gerrymandering, the nasty process where political districts (mostly in the USA) are drawn by partisans to allocate voters and increase the probability their party will win seats in congress, state legislatures and other bodies. It's cheating, but all attempts to stop it so far have failed. It got worse in 2010 with the Republican Party's "Operation Redmap" where they put a special effort on winning state governments to give them redistricting power. It has been described as "Instead of voters picking their representatives, the representatives pick their voters."

It's been hard for the supreme court to fix this, even though they have said they wish to. It's not their job to write districting rules, they can only tell you if the rules you wrote fit the constitution and the law. Little hope remains there.

Here's how to fix it, with a special multi-state "agreement." I write agreement in quotes because while it would be best to do this as a binding interstate compact, that requires the consent of congress, including the Republican Senate, and may not be feasible. But it's possible to do it without such consent in a different way.

States joining the pact would agree to:

• Draw their districts in an algorithmic and non-partisan manner. There's software now to do this, keeping districts compact, concave but also in compliance with the voting rights act. They also delay their redistricting deadline to the latest possible date.
• Track the districting of states which didn't join the pact. Use the algorithms to calculate how gerrymandered they are compared to how they would be if they joined.
• Whatever bias is found in how non-member states drew their lines, the member states instead redraw their lines with an algorithm modified to gerrymander their districts in the opposite direction, plus one for each state.

The basic result -- any state which doesn't follow the rules of the pact and biases their districts causes their party to net lose a seat in the House of Representatives. If a group of 3 Republican states gather to bias seats by +10 Republican, the states in the pact counter by drawing their districts +13 Democrat, for a net bias of +3 Democrat. In the end, gerrymandering becomes not just pointless but counterproductive -- it only makes your party lose, not gain seats. (It does protect individual members though, which is a concern.)

In a rational world, it ends. The states all agree to the pact, and draw their districts in a non-partisan way from then on.

Complexity

As noted, it would be nice if the pact were a legal "Interstate Compact" which is a binding agreement between states, agreed to by congress. If congress won't agree, the agreement can still exist, just not sued over in the courts. In particular, as long as the pact has enough loyal members, enforcement is not really needed. If a state breaks their word, they get punished, not through the courts, but directly, by having their actions reversed like a non-member. In addition, member states can exact any other penalties they can legally apply. While states can't engage in trade wars, they have a number of legal tricks up their sleeves they can use to make it hard for states that break their word, particularly the economic powerhouses like California and New York. They would not apply these punishments for just not joining the pact, only for lying and breaking their word.

In addition, membership in the pact would require enabling legislation which can be written in a way it's hard to break, including requiring a super-majority to cancel. In that case, suit could be brought within a state if they break the pact, by suitably minded citizens of the state.

Non-membership

For a non-member, they would be free to district as usual. After they drew their districts in some way that has finality, the calculation would be done. However, if a state deliberately delayed their drawing until the last possible date, then a special "worst case" calculation would be done. In that case, individuals could show extremely gerrymandered district borders which are the most extreme. For a state that waited to reveal, it would just be presumed they were going to do the worst, and the penalty would be calculated based on that. Few states actually do the worst, so it would be worthwhile for them to get their districts down earlier for the lesser penalty.

Non-member states could draw their districts in non-algorithmic ways, as long as the net result (based on analysis of voting records from the most recent elections) is the same as the algorithmic approach. Indeed, even member states could do this but it would be discouraged. No penalty for that.

Not even a pact

In a way, this could be treated as states acting individually. There is no strict need for states to coordinate, though it's useful if they work together to do the counter-gerrymander. California on its own, however, could eliminate its 8 Republican districts. California currently has non-partisan districting. Add in a few other states and even without a pact, but just with a common frame of mind, they could make this happen.

So how much Gerrymandering is there?

It turns out that the USA is "naturally" Gerrymandered. That is to say, if you use a fairly simple compact district algorithm, it tends to give about 26 extra seats to the GOP than the Democrats, even though the Democrats have more voters! This is a consequence of the population densities and the fact that Republicans concentrate in rural districts while Democrats concentrate in urban districts. It gets a little better when you add the "majority minority" rule in the voting rights act. This act demands that some districts have a majority of the voters be members of a minority, to assure there are some minority chosen representatives in the House.

I strongly recommend that all read the results of 538's Gerrymandering Project for some surprising details.

Experiments show you can also draw districts on a "most contested" algorithm that tries to make as many districts be "swing" districts, while leaving the rest safe. But some are just going to be safe no matter what you do, and more of those are Republican unless you go to strong measures to stop it. What the best formula for districting is could be the subject of much debate, and will be. This system assures that once the system is decided, states can't deviate from it to help their party. While I would like a system where the eventual distribution of the House matches the allocation of voters state-wide or nation-wide, I am wary of deliberately drawing lines for even anti-partisan rules rather than non-partisan rules. The states in the pact will have to decide how to do that, and hopefully they will do it fairly.

In addition, lines should be redrawn frequently if possible, to adjust for changing population, if they are to be based on voting patterns, which change quickly.

The only way to actually remove Gerrymandering would be with some sort of proportional representation system, which typically involves not having districts per se. Current federal law requires there be contiguous districts, but it could be changed, but not using the process above.

Covetous Representatives

Redistricting puts House members into panic. Lines are moved, districts are sometimes merged or eliminated. That means some incumbent loses a job. The switch to algorithmic, non-partisan lines will mean a lot of incumbents lose their jobs. I think that's good, but they won't see it that way. This is one reason it's good to not worry about the consent of Congress. Even so, reps will pressure their states against doing this for their own selfish reasons. This must be guarded against, but that's difficult to do.

States all have different rules for when they draw the lines. Because member states will just use an algorithm, they don't have to do the formal line certification until the exact deadline. Of course, people will still be able to see the expected result as soon as census data comes in, though it will change based on what non-member states do, if they don't give in and do the smart thing.

The hard deadline is that districts must be drawn before the first primary, which is generally the spring of 2022. Some states have other deadlines, which are outlined here. However, once a state passes a law they can't repeal (because it requires consent of both parties) their lines can be considered done.