Adumbration

The Chicago Tribune alerts me to news in Illinois to improve the situation surrounding a topic I alluded to a few months ago - gerrymandering.  Gerrymandering allows back room political deals to control political borders that protect special interests and their politician tools.  The Tribune's John Chase writes:

"Advocates trying to change the way Illinois political maps are drawn began the arduous process today of trying to prove they have enough signatures to get a constitutional amendment placed on the ballot this November."

Evidently there is some controversy regarding how much time has been allowed for the initiative sponsor to correct what the elections committee has provisionally declared to be invalid signatures. Maybe not surprising that agents of the state would try to roadblock the reform of one of the tools politicos use to remain in power, so I became interested in the organization sponsoring the amendment - "Yes! for independent Maps."

I support change that tends to reduce the concentrations of power, power that allows politicians to more easily entrench, but we have seen many cases where promises to reduce the scope of government led instead to even more of it (e.g. the Reagan presidency).   So the actual conditions of the proposed amendment are of some interest.

The Tribune gives a fair summary, but the few additional details in the amendment wording might be worth analyzing in greater detail.  To that end, here are the few sentences that describe how district boundaries are intended to be established:

Legislative Districts shall be contiguous and substantially equal in population. Representative Districts shall be contiguous and substantially equal in population.The district plan shall comply with federal law.

Hard to find any fault with this statement; one would expect a republican form of government might make these districts target equal numbers of people.  To "comply with federal law" might be redundant given the primacy clause.

Subject to the foregoing, the Commission shall apply the following criteria: (1) the district plan shall not dilute or diminish the ability of a racial or language minority community to elect the candidates of its choice, including when voting in concert with other persons;

Much might get hidden in these words over time - I think the basis of presumed community has often been used to justify the salamanderish routing of political boundaries that maintain the power of one or another interest group.  The final clause could be interpreted to restrict such ability - we are not any of us only one thing, or only some other thing like "white" or "black", "worker" or "entrepreneur," "mod" or "rocker", and to that degree will often or always align vote "in concert with other persons."

(2) districts shall respect the geographic integrity of units of local government;(3) districts shall respect the geographic integrity of communities sharing common social and economic interests, which do not include relationships with political parties or candidates for office; and

These clauses might be aiming to throw a bone to local politicos, and people aiming to manipulate this amendment could twist the notion of "common social and economic interests" into pretty much current status quo.  But I like the final clause, and the wording should make this the dominant clause as it controls and limits the others. I am concerned with the looseness of the word "relationship", but in this context it might serve to expand the range of meaning and make it more difficult for political interests to sneak in.

(4) the district plan shall not either purposefully or significantly discriminate against or favor any political party or group.

This wording might also be interpreted to maintain current interests, in as much as any reduction of the currently twisted system would diminish the power of office holders of any political party.  But the wording seems careful, maybe intending to avoid that sort of interpretation.  It seems that proving actual favor to a political party or group would be fairly easy to accomplish, which as a low bar would make it possible for small groups of determined plaintiffs to challenge any particular plan.  That, in turn, would encourage the planners to take pains to avoid even the appearance of impropriety.  There are features of the planner selection, and process transparency, which also encourage compliance with what seems like the spirit of this amendment; perhaps I will describe those in a future posting.

In designing the district plan, the Commission shall consider party registration and voting history data only to assess compliance with the foregoing criteria, and shall not consider the residence of any person.

This clause would be difficult to verify, as we can't with certainty tell what some person is considering when they make recommendations or decide on voting alternatives.  But it should be possible to perform statistical analyses of the plan as compared to the voting history and residence, which would expose the obvious violations - we should expect no statistical correlation, as any violation of that would strongly imply a bias introduced by the people involved in the process, given that the process is performed by human beings.

Which segues into an outline for a further improvement to settling voting borders (short of getting rid of them completely, a subject for another future post).  Such improvement could be an impersonal algorithm driven by population density and equal population-weighted geometric paths.  Picture a map that highlights population density, where islands of high population density rise above the ground surface, like skyscrapers climb over a city.  Across a given state, areas around cities would appear higher in this view, and those points would serve as the centroids of the districts - the highest N points in population density corresponding to the population centers of the N districts.  The area through which population is averaged will make a difference in the implementation, but could be chosen on the basis of the largest computation size that would still differentiate among N peaks, and once established on even a random basis would thereafter be relatively unbiased.  Note that the population averaging size that is used to determine the N centroids need not be the same (and most likely would not be the same) as that required for the districting plan.

Once the N points are established, one projects from each point in all directions at a rate proportional to equal population density, until each cell touches one other cell in each direction, which should be the population mid-point along each line between the N points.    If any islands arise with population through this method being substantially different than that of the other islands, then reduce the population averaging size.