American style democracy has had a great run and probably has a few generations left in it. But there are clearly some signs of age we’re running across, and things that should probably be fixed. If we could start over today, what would we do differently?
In this, part one, I’ll talk about some potential fixes to gerrymandering. Gerrymandering is what happens when you put the same people who you elect in charge of how elections get done. Districts get redrawn in such ways that the parties that control those districts are less likely to lose them.
What this does is what a lot of methods I’ll talk about do: make a politician loyal to her constituency. A politician’s constituency ought to be her boss, however, with heavily gerrymandered districts, a politician’s boss often becomes her own party. She’d better toe the line, lest she risk losing national funding for her campaign and be replaced with someone more loyal from inside the party.
There are many proposed fixes to gerrymandering and two dominant schools of thought. One is that independent commissions ought to be installed, divorced from electoral politics, and consisting of members of both dominant parties as well as independents. These commissions can oversee redistricting as they see fit.
The other dominant approach is to use algorithms to redistrict along ‘most compact’ lines. This removes any and all political bias and enshrines the method by which districts will be drawn.
I take a cue from our own constitutional democracy, and propose a combination of the two, similar to how federal districting in Mexico works.
Algorithms, as the link describes, can be made to be NP-Hard if complex enough. The simple algorithms described to create ‘most compact’ lines leave out a lot of details that we’d want in any real districting, such as ensuring we don’t accidentally redistrict out a protected minority. But adding in these details can me it’s almost impossible to get the ‘best’ answer out of an algorithm – it may take modern supercomputers until the end of the universe to best calculate the ‘ideal’ districts.
And so, why don’t we take a compromise approach? NP-Hard problems often have heuristic ‘cheats’ that will give you an arbitrarily ‘good enough’ answer in a reasonable amount of time. So we’ll start there – get some ideal districts from the algorithm.
The second step is to send these districts off to the commissions. I don’t like the idea of having commissions made up of the two parties and some independents – the two party system and ‘vote for the lesser evil’ is partly what we want to fix, especially since it gives a huge incentive to spend large amounts of money on negative advertising. Instead, commissions should be larger tent committees with at least any party that can get 5% of the vote in a national election. There may be more inclusive and liberal means to define who gets on a commission that we can re-evaluate in time.
And so, to end gerrymandering – redraw the districts using standard, public, transparent algorithms. Do any and all remaining horse trading to finalize districts with very politically diverse committees. This should help hold a politician loyal and responsible to her constituency, rather than to her party.