29 December 2010

How to prevent Gerrymandering

In the wake of the latest Census reports comes the usual hand-wringing about the drawing up of congressional districts within each state. Ever since the emergence of permanent political parties, the power of partisan legislative majorities to divide a state map into monstrously shaped districts -- the original "Gerrymander" was a cartoonist's anthropomorphism of one such district in Massachusetts -- has been deplored by generations of idealists who've hoped and prayed for a rational, nonpartisan way to portion out representation. It seems not to have occurred to people that the original problem is in the Constitution itself. The Framers envisioned the House of Representatives to be the "people's" house of Congress, while the Senate represented the states. Yet once the federal government determines how many representatives each states should have, it's up to the states to draw districts. But if the House is meant to represent the people, and not the states directly, why are states part of the process at all? Why does the federal government divide the magic number of 438 among the states at all? The fact that the number of Representatives is set and only rarely altered may also be part of the problem. The solution may involve giving up the idea of having the same number of Representatives in each Congress. It should also involve skipping the states, since the House isn't their branch, and basing representation on the next most local level of organization: counties or their counterparts. These counties are pre-existing units; they aren't constructed arbitrarily for partisan purposes. Assigning representatives to counties eliminates gerrymandering. Let the federal government define the minimum number of people entitled to representation. If a populous county has more than that number, they get the appropriate number of representatives. If a sparsely populated county falls short, let a rule be established for bundling together contiguous counties until the combined population reaches the minimum for representation. If that means merging a "red" and "blue" county for election purposes, so be it. This plan seems simple enough, and I can't immediately imagine objections from either major party. States-righters might complain about a loss of sovereignty, but we've established that the House of Representatives isn't supposed to represent the states in the same way the Senate was meant to, so they have no cause for complaint. In any event, it isn't as if some distant federal authority would be dividing states into districts on an arbitrary basis; that's been the states' business all along. If people are serious about ending gerrymandering, and aren't just making a show of their highmindedness, they'd get to work on drafting and promoting a constitutional amendment to correct the problem. As I've pointed out, apart from the tedium of the ratification process, this shouldn't be that hard.


Dale Sheldon-Hess said...

Two quick points:

1.) Its 435, not 438 (you're conflating the 3 electoral votes for DC with house seats.)

2.) Not all states have counties (AK), and some that do have no county-level government (CT). Not insurmountable, but a complication.

But I also have a larger point: simply asserting that "some rule" should be established for giving multiple seats to some populous counties and combining some less-populous ones shows--and forgive me for saying this in such a blunt way, but I fear it must be done--a real ignorance of how this problem has been grappled with in the past.

Are you familiar with WHY it was decided to set a fixed-number of seats? It used to be that, every 10 years, there would be HUGE arguments over how many seats should be in the new congress. These were caused by legitimate claims of injustice, because there are several mathematical paradoxes that occur when you allow the number of seats to vary and when different apportionment rules are used. Things like "increasing the total number of seats causes a state to lose a seat". These arguments are what caused 1920 to be skipped COMPLETELY for reapportionment: congress was completely unable to come to an agreement as to how many seats there should be, and how they should be divided, a clear failure of a constitutional mandate.

I recommend "Numbers Rule", which is primarily about the history single-winner voting systems, but also has a handful of chapters about precisely why this problem is much harder than you seem to think.

Samuel Wilson said...

Dale, I appreciate the correction and the information. Some of your objections from history are irrelevant to my concerns. I propose not to care, for instance, whether or not a state loses or gains seats since the state isn't directly represented in the House except when delegations must vote as states to decide a presidential election. That rule, too, can be changed. The real question raised by your criticism is whether states' interests are entitled to consideration when seats in "the People's House" are distributed. If no Representative is answerable to his state government, the answer would seem to be no. Since Representatives are meant to serve constituencies smaller than states, I feel entitled to ignore the complaints of states across history. If objections have been raised by actors other than states, I'd appreciate the clarification. Also, should gerrymandering not pose a problem for you,-- I can't presume one way or the other from your comments -- you're free to dismiss my suggestions completely.

Dale Sheldon-Hess said...

Surely, the people IN each state had a concern for the number of house seats their state had; you must admit there is at least some state-pride and some shared desire among citizens of a state. That their elected state leaders and federal representatives share that enthusiasm shouldn't be a surprise. I'm sympathetic to your feelings; what's less than a quarter of a percent in the grand scheme? But people--not just their representatives who would be out of a job, but people--are sometimes quite adamant on these things.

I am opposed to gerrymandering (I think any sane person would be), but I'm also familiar with the research indicating that it's nowhere near as big a deal as most people think; under any districting scheme, very nearly every election ever in this country would have fallen along the same partisan lines (third parties would not have done any better, and only a few seats in any congress would have gone to a different major party; it's noise.)

I quite like some of the algorithmic/mechanical districting methods out there, but I focus my efforts for systematic change elsewhere (campaign finance and approval or score voting.)

Samuel Wilson said...

Dale: I agree with you that districting reform isn't the highest priority compared to ballot reform and relief from the tyranny of fundraising. My original post was just a modest proposal in response to the decennial griping about gerrymandering or the concoction of "majority-minority" districts. Democrats in particular complain that particular practice, but I'd be interested in seeing scholarly work that attempts to debunk their complaints. It may be that criticism of gerrymandering is only so much mugwumpery, but if critics really want to do something about it, state pride will have to be confronted on some level, and local pride cultivated.

Crhymethinc said...

Another alternative, of course, would be to eliminate states and their governments completely.

I, for one, would much rather see "one nation, under..." rather than 50 squabbling states, constantly backstabbing one another to curry favor with big business and special interests.

Samuel Wilson said...

Crhymethinc: "One nation" is a fine notion, but the lower house of Congress would still presumably include representatives of specific geographic units. Would Congress itself then divide the entire country into counties or districts? The challenge would remain a matter of preventing partisans from drawing the borders to maximize their competitive advantage in congressional elections -- presuming that you believe that gerrymandering can make a significant difference in the first place.