Digital Analogue

A modern "districting"

Here's a totally wild idea I wanted to throw out there. Right now most jurisdictions have some sort of geographically-based districting, with each district electing some number of representatives. Other common systems include at-large representation (all representatives elected by the entire jurisdiction) or party list representation (where each party submits a list of candidates, which are then "elected" in proportion to that party's share of the vote---this is common in Europe).

All of these methods have a tendency to marginalise certain voters. What if people could change the district they were in?

A bit of thought dragged this idea to its logical conclusion, which actually is relatively robust, I think. Each jurisdiction has, instead of N districts, N representative groups. Each group elects one representative. Every person belongs to exactly one representative group, of their choosing. People can change groups at any time (possibly with some logistical restriction on frequency of changing). That's it! No further restrictions.

The first objection I came up with: what about keeping districts the same size? Totally not a problem. If the system gets a rep group with too few people and another with too many, the one with too many can easily migrate half of its people over to the small group, effectively taking it over.

So what about the small groups that get taken over? They build coalitions in order to securely maintain one rep group.

With respect to third parties, this sort of system would enable them to form more easily as a rep group of like-minded people that can successfully elect a representative from that party.

With respect to existing parties, there would undoubtedly be a significant number of rep groups that map fairly well to the existing parties; but well-known fault lines would probably be reflected by having e.g. one of the Democrat rep groups be anti-choice.

Minorities of race, gender, and culture that felt it important to be represented by one of their own could easily form a rep group to make it happen.

What about the things the old geodistricting was good at? Rural districts really do sometimes share concerns that the urban ones don't; and there would almost certainly be many rep groups that were dominantly one or the other.

Essentially, the system is maximally flexible to reflect the up-to-date current political fault lines. It's in a representative's best interest to represent their constituency well, because otherwise they will drift off to other rep groups, leaving this one ripe for a takeover by some other group with their own candidate. Or maybe they'll stay here and just elect someone better. When not actively performing a takeover, it's in a citizen's best interest to join the rep group with the candidate most closely aligned to them (both in outlook and priorities), because then they can help to protect that representative from a rep group takeover and exert some influence over them as well.

In terms of implementation, the chief difficulties are in keeping track of membership of the rep groups and in managing elections. The former is only incrementally harder than maintaining voter rolls now. The latter is a little trickier, but not hard if we move to machines that actually print the ballot in addition to helping you vote---the machine can print the ballot appropriate to your rep group(s).

So, there it is. Neat idea, huh?

"'No Scrubs,' TLC. It begins with a definition. It has axioms. It makes inferences. How cool is that?!" --Annemarie Peil