With the 2010 census and resulting redistricting coming up faster than we're liable to be ready for, I was thinking about getting involved in efforts to make the redistricting process less partisan (which I take as de facto a "good thing"). Not wanting to go rushing into this uninformed, I've done some recent reading. In particular, I found this report from the Brennan Center to be thorough, balanced, and very helpful.
Below the fold, I bring up some specific issues relating to redistricting that now seem less obvious to me than they once did. I'm hopeful that your comments and your poll responses will, at a minimum, inform me of the progressive point-of-view on this topic. Thanks!
Let's start off with the proposition that there are actually two ways to make the redistricting process less partisan - 1) turn it over to a bipartisan/nonpartisan group, or 2) create rules that severely limit the flexibility of whoever the line-drawers may be.
It does seem to me that the unavoidably partisan legislature should be involved beyond a mere up or down vote. Inspired by the Brennan paper, I'm inclined to believe they should have some (constrained) tweaking power. Plus, I acknowledge that finding a truly nonpartisan group - given the enormity of the stakes involved - is going to be essentially impossible. Consequently, I think we need to rely upon rules.
As an aside, I admit that I'm not confident about putting into place rules that can't be ignored. After all, any rules that are enacted as mere statutes can certainly be effectively amended/repealed by simply passing legislation that adopts a redistricting plan which ignores those rules. For instance, if we have a law that no one may marry unless they have black hair and then later pass a law laying out the specifics by which redheads may obtain a marriage license, redheads are going to marry. At the very least, they'll have a damn fine court case. One is left to wonder if constraints on the state legislative redistricting process must be enshrined in state constitutions in order to be effective. (I note that a simple federal statute would be effective with respect to Congressional redistricting.)
So then, it's rules. Ah, but what rules?
There are a couple obvious starting points, although even those are somewhat less clear than we may think.
We start certainly with "one man, one vote". Equality of population is one obvious goal, but how equal? With 5,000,000 people in the state of Jefferson and 100 legislator in the lower house, must there be exactly 50,000 inhabitants in each district? It seems that SCOTUS would allow state legislative variations in the range of 45,000 to 55,000. Is that too large? Then what is the appropriate degree of equality, particularly when dealing with a moving target? The equality of each district will already have changed by the time the new districts are first implemented and certainly will change in the ensuing decade. Plus, the greater the emphasis on equality, the more other desirable redistricting goals may have to suffer.
The other obvious starting point is compliance with the 1965 Voting Rights Act. I'm sure you're all sufficiently familiar with that legislation. In any case, I'll touch at least briefly on this again a bit later.
Even with the two constraints above, line-drawers have an enormous freedom. Where best do we start to rein them in?
I think there's fairly broad agreement that "gerrymandering is bad". (We're going to proceed from here in a hypothetical environment: whatever rules are adopted will apply everywhere. This is not a case of California agreeing to split up their electoral votes while Texas continues as winner-take-all. There is no cross-state partisan advantage to be had.) Rules typically used to limit the excesses of gerrymandering place restrictions such as: contiguity, compactness, respecting pre-existing boundaries of cities & counties, and not dividing communities of interest.
I don't think anyone has any serious objection to requiring contiguous districts. On a gut level it simply makes sense. However, it's generally required now and we know it hasn't exactly stifled the art of gerrymandering (insert your own favorite example here:_____ or check out this one)
There are reasons to like existing boundaries, like county lines. It does somewhat simplify the "political map" which may make it easier for the voter to know who represents him/her (e.g., "if I'm in Podunk City, I must be in District 1"). It also makes ballot design simpler which makes election administration slightly less prone to human error. Plus, if you keep entire towns/cities/counties in single districts, perhaps that tends to keep together "communities of interest", as it obviously would for urban/rural concerns. And existing boundaries tend to be at least somewhat compact (although the detail level boundaries of cities may be very erratic, often the over-all result is reasonably compact). Perhaps this aspect deserves a high priority in line-drawing?
But what about measuring compactness directly? There are formulas to do so, shall we mandate some limit? Require some average? After all, gerrymandering can be in the eye of the beholder, so maybe a mechanicial formula is best? But on the other hand, it doesn't take a genius to design equal population districts that are nicely compact but still manage to gerrymander using the classic techniques of "packing" (stuff all the minority group into a single district thus presumably limiting their influence to that single district) and "cracking" (divide the minority group's geographic area into multiple districts and thus presumably (but reference the Obama win...) making all their voters ineffective at electing one of their own).
And what about "communities of interest"? For starters, how do we define them? How do we know which communities should be considered and which not? And what kinds of redistricting actually best serve their interest? After all, consider "majority minority districts", that is, districts designed specifically to (virtually) ensure that a candidate who is a minority of the whole population is elected from that particular district. Arguably, this is an instance of "packing", a classic gerrymandering ploy intended to under-represent the group in question (the other classic being "cracking"). Plus, this can result in districts that are - at first glance - obvious and clearly undesirable gerrymanders (for example see FL03 which is a classic majority minority district) despite being almost semi-required by the Voting Rights Act. It seems to me, anyway, that "communities of interest" are a concept too vague to be given weight in the redistricting process. However, going back to the legislature's "final tweak", maybe they're appropriately added in at the final stage?
Of course then there's the whole question of partisan impact. Unless the process is entirely in the hands of true nonpartisans who have little personal knowledge of the interplay of their state's partisanship and geography, nothing is likely to influence the final map more than which party will benefit the most. Considerations like boundaries or compactness or communities of interest or anything else will play a barely heard second fiddle.
In an alternative reality, I might be tempted to allow the politicians of the state of Jefferson to divide the state up into 80 house districts of essentially equal population in whatever way they wanted, limited only by the requirement of contiguity and some very forgiving definition of compactness. But then I'd use the remaining 20 seats to "true up" the total number of seats in the legislature to more closely reflect the actual percentage of votes cast throughout Jefferson. But sadly, kind reader, neither you nor I live in that reality.
In our reality, this aspect is the number one reason why the hammer of restrictive redistricting rules must come down hard. And our very first question here has to be: Do we eliminate/mute the partisan impact by ignoring it or grab the bull by the horns and explicitly deal with it?
Let's say we mandate a priority list like this:
#1 Equal population +/- 0.25%
#2 Compliance with Voting Rights Act
#3 Follow existing City & County lines except to the extent required for -
#4 Compactness within some semi-tight maximum value for each individual district as well as a strict average compactness over all districts.
If the above list is found to give too much "rope" to the line-drawers, go back and make the rules stricter. I'm no expert, but I'm going to guess that even with rediculously strict rules, there are going to be opportunities for experts to get the results they like in terms of partisan impact. Consquently, the answer to the above question about reducing partisan impact is to deal with it explicitly.
Let's say that our dearly beloved state of Jefferson has the following breakdown of voter registration (yes, I know that not all states register voters by political parties, but most states do and so does Jefferson):
Dems 41%
Reps 37%
Other 1%
Indies 21%
Now re-allocate the Indies in some fashion (argue/adjust/refine the details later; simply accept for now that a fair method can be found). Maybe you end up with:
Dems 52%
Reps 47%
Other 1%
Or expressed in terms of the 100 member lower house of Jefferson:
Dems 52 legislators
Reps 47 legislators
Other 1 legislator
Let's also say we've got the above 4-point priority list.
Once a map is drawn, the voter registration stats for each new district are compiled. For each district, a statistic is calculated. I don't have a specific equation, but it has to make some sense in terms of the likely electoral impact. Let's say that if the two leading parties have equal registrations, then the statistic is 0.5 for the Dems and 0.5 for the Reps for that district. If one party has a registration lead of 10% or more, the statistic is 1.0 for that party. In between? Again, just trust that something suitable/fair/sensible can be arrived at (and I realize the complexities added by including "other", but c'est la vie). Sum up the 100 statistics. If the summed statistic varies by more than some prescribed amount from the above derived proportional representation, the map has to go back to be re-drawn no matter how thoroughly the 4-point list is satisfied.
Again, I admit I'm early in my research on this. Surely there are things I'm missing (if so, I'm glad to have you folks around who will definitely set me straight, and thank you for that). But at least at first blush, this final 4-point list with a partisan balance requirement seems capable of giving us rational districts that can meet special needs without giving unwarranted political advantage.
Looking forward to reading your comments. I apologize in advance that I've got to get offline and won't have much chance to participate in discussion.