Slashdot Mirror
Redistricting Videogame Shows Problems in the System
An anonymous reader writes "This is a cool redistricting game that was launched out of the capitol building in Washington DC last week. It was created by the USC Game Innovation Lab and has been getting lots of press. It's about time someone took on a tough issue like redistricting reform using the power of the internet." It's crazy that gerrymandering is actually good fodder for a video game.
Or did you think that American politics at the highest levels was actually about serving the public?
"Trolls they were, but filled with the evil will of their master: a fell race..." -- J.R.R. Tolkien on Olog-hai
One of the big issues in redistricting is minority representation (or non-representation), which leads to districts that consist of urban regions connected by a thin corridor or other similarly bogus shapes. Instead of artificially trying to group minorities (or party strongholds, or whatever) into specific geographical areas, though, why not remove that layer and replace it with a system that inherently represents various groups proportionally?
Using a single transferable vote system like that used for Cambridge (MA) municipal elections could work quite well. In the city council race, there are 9 seats, and any group capable of generating at least 10% of the total votes can elect a councillor of their own, even if that group is spread from one end of Cambridge to the other. Some councillors do have unofficial "districts" where their support is strongest, but this is not a requirement in any way.
STV elections also avoid the "wasted vote" problem with independent or smaller-party candidates, since voters can put one of those as their #1 choice, and if they don't win, those votes transfer down the ballot to the #2 or later choice as necessary.
With the current breakdown of seats by state, a system with a maximum of 11 seats in a district would allow all but 11 states to operate as one large multi-member district; raising the threshold to 13 would add Georgia, New Jersey, and North Carolina to the single-election list.
To use Massachusetts as an example: the current 10 seats in the House are all held by the Democratic Party. I doubt there's any viable redistricting that would allow the Republicans to win even one seat. Under a 10 member STV system, though, the 13% of the state that's registered Republican could elect at least one, and with support from unenrolled voters, possibly more.
The best system for districting the US seems to me to be the one based on post offices. Each post office does define a community, especially in Federal services terms. It serves a small group of people who live very close, sharing mostly the same conditions other than those inside their private dwellings - which are also likely to be similar (and even homeless locals have the same access). It is the most common face of the Federal government, directly serving the community. And it already services election procedures like registration and delivery of election info.
I like the system where each person in a post office's service area (usually a ZIP code or two) selects the neighboring postal zones (up to the state border) to which they're most "connected" in order of "closeness" (as defined by the person selecting). Then all the responses are tabulated purely statistically to generate a map of the most interconnected regions, in a quantity equal to the number of representatives allowed in the state. There could be a second round to accommodate exceptions, like tiny islands (below some predetermined population size) or extremes of minimum/maximum populations in different districts, where the exceptional zones select their associations, as do the neighboring candidates for association to accept association with the exceptional zones.
We should choose our own fellow constituents who choose our mutual representatives. As long as the politicians themselves mediate the process with any discretion, the process will primarily serve them and their parties or other interest groups. We've got the stats and the sense of our neighbors to do it equitably and quickly. We should redistrict at least 10-20% of districts every odd-numbered year for reelection to the House of Representatives on the following year. After no more than a decade or two we should have equitable districts without a hasty conversion that will generate unmanageable sabotage from the existing order.
--
make install -not war
What you're not taking into account is that usually the change in majority comes only from a major, major shift in public perception of the current bunch of weasels, faster than they can compensate for with redistricting. E.g., last Congressional election, and the "Republican Revolution" back in the 90s. And this last time, the new majority party just barely managed to squeak through with a majority. I don't recall how much the Republicans won back in the 90s, but I know that the election immediately following it had them just barely keeping their majority.
The congressional incumbancy rate was 98% in 2000.
To my thinking, the solution is simple: mandate convexity of the districts, with an exception for irregular district borders at state boundaries. Districting would then become a sort of Voronoi diagram over a non-uniform space due to population density. This would reduce the problem to one of choosing the centroids of each district, which would be much harder to manipulate inappropriately due to the complexity of the problem. Still, you could define the locations of the centroids based on some metric such as maximization of distance between the centroids.
I was thinking about this idea a few months ago: You create a realistic sim-type game that when played, it encourages the player learn or develop a particular political point-of-view, simply by demonstrating how things work or don't work together. There was an old game from the mid-late 80's that sort of worked that way called Spheres of Influence.
BTM
That was the turning point of my life--I went from negative zero to positive zero.
http://bolson.org/dist/
I think I've gotten pretty good results for CA, TX, IL, FL and PA
It tries to create impartial districts that keep people on average close to the center of their districts. It works pretty well, but is kinda computationally intense. It could almost become Redistricting@Home if there was interest in the approach.
Start Running Better Polls
I am in fact familiar with the concept of money...
The point here is social inequality. Money may be a proximate factor for social inequality, but I find it unlikely to be the ultimate cause. Money has been around a long time, and throughout that time there have been wide swings in social inequality - it seems desirable to minimize inequality, while maximizing both average and total wealth.
There are other ways to keep those born rich from staying rich, besides your modest proposal, such as inheritance and progressive taxes. These things exist today for that very purpose, but somebody's going to call any type of tax unfair. I think we could stand to be a tad more progressive, making things a little easier for the poor and a little harder for the rich and the corporation, especially in these days of increasingly consolidated wealth, shrinking middle class, and growing poverty.
Before anyone calls me a pinko commie, I think that that the promise of personal wealth is the greatest part of capitalism. However, the grubby capitalistic hand needs to be slapped from time to time to keep it from harding everybody else's cookies. Besides, while money necessarily provides an advantage, there is no reason why that advantage should pay such high dividends to the rich due mainly to its interest bearing nature, and yield such low returns to the poor.