The absurdity of US redistricting

Nate Silver runs down the likely impact of the newly released population distribution on how the House districts may be redrawn before 2012.

What an absurd electoral system.

28 thoughts on “The absurdity of US redistricting

  1. I realise that it is at the end of the re-apportionment cycle, but the largest district is nearly twice the size of the smallest?! I didn’t realise that it was that bad in the House.

  2. Federations seem to have a problem when the size of the legislature is set to a fixed number. The NSW/Tasmania ratio is a lot less than the California/Wyoming ratio, but even so the variance between districts with the largest and smallest enrolment is between the ACT and the Northern Territory at 120875/59207.

    That may be a good test case because ti is not effected by any guaranteed seat minimums. The obvious solution is to determine the number of seats per region by applying a uniform quota to the regional populations.

    Just for the record original states in Australia are guaranteed a minimum 5 seats, but that does not increase when the size of the House increases. I believe the apportionment rules in Canada are worse.

  3. I also note that some high-population states have ‘unbalanced’ districts, presumably driven by a 10-year cycle, bolstered by an inability to take into account expected future relative population changes (both between and within states). I see NC-9 is ~42% larger than NC-1.

  4. Ed, what particular aspects are you asking about? I think that the must-be-as-close-as-poss-to-census aspect is the worst part, but there might be aspects I haven’t thought of.
    Australia (also a Federal country) has
    this method.

  5. it is more than faintly weird that the US has one of the shortest legislative terms in the world, to keep the legislators accountable to the people, but one of the longest redistribution cycles, presumably to keep the legislators not accountable to the people.

    The UK, for reasons I do not understand, not only takes decades to complete redistributions but uses census figures from several years before the time they commence.

  6. Yet the UK increases the number of seats in the Commons for almost every election.

    I have been meaning to put up something on Canada’s reapportionment, as it was in the news there recently.

  7. Errol, I am not aware of another country that has legislators determine the boundaries (or magnitudes, where relevant) of districts.

    That’s not to say that everyone uses neutral civil servants or genuinely independent commissions as their exclusive means of doing it. As I alluded to above, a bill in Canada that would have changed the reapportionment (although not actually drawn the district lines) was recently pulled back because of the partisan controversy it generated. And a plan to change the process in the UK was tied to the call for the referendum on AV (though I missed whether the two issues were ever decoupled).

    There is a big difference, however, in these cases of politicians tinkering with the process for partisan advantage, and politicians actually drawing district lines themselves. So, the US may be unique in this regard.

  8. The reason the House of Commons increases by several seats at nearly ever election (absent a conscious cutback as the Cleggemony is now proposing) is that the Rep’n of the People Act requires the Commissioners to use the average population per seat from the current House as the quotient to which districts must, “as nearly as practicable”, approximate. This means they end up using that average as a d’Hondt quotient since, all else being equal, more seats rather than fewer (7 rather than 6 if a county has 6.25 times the quotient in population for example) is more likely to equalise disparities among individual constituencies.

  9. The UK may, just may have a sillier system of redistribution than the US.

    There are different quotas for each country in the UK. There is an extraordinary 25% margin that seats can depart from the quota in their country. The process is based on actual enrolment at the start of a review. A review can take up to 8 years and is only required every 8 to 12 years. The standards are inconsistent and there are in any case about a bazillion special circumstances exceptions to what standards there are. The separate boundary commissions for each country are at least neutral bodies.

    It is really quite hard to understand why no-one seems to be suggesting a uniform quota, coherent standards and a tighter margin.

  10. The New York Times article on redistricting today mentioned that in the US, the politicians chose their voters, and implied in a real democracy it would be the other way round.

  11. I do not quite get how the house officials could propose to have the largest district twice the size of the smallest one… Whatever reasons they invent and feed to the public, this seems like a clear cut example of gerrymandering to me. I wonder what measures could be taken to prevent such things from happening again and again.

  12. You might be interested in


    The system was devised in the 18th century. If it had been left open to legislation, no doubt they’d have stopped redistricting sometime in the late 19th century and a redistricting now would be unimaginable radicalism, probably Communist in nature.

    Still, you make a good point and we should probably shorten the period by amendment — maybe tie it to the Presidential cycle.

    Oh yeah, merry Christmas, happy holidays, and a very nice Yule, to all.

  13. My mistake re UQ – harmonic mean as opposed to St-Lague cutoff for rounding up fractions of a quota, not d’Hondt.

  14. Aaron

    I think much larger legislative bodies are a very good idea. I note the Cameron-Clegg government is at least moving towards a uniform quota, many fewer exceptions, and a faster cycle.

    Ideally Australia and the US would move towards a uniform quota based on the Shugart number, which would then be applied to the population of each state and territory. The minimum representation for a state would be 3 (more than one for proportional representation, an odd number to avoid Tullymandering) It seems to me the ideal cycle is South Australia which redistributes after every general election.

    The redistribution body, ideally the election management body itself, should be neutral, professional and nonpartisan. NB Nonpartisan does not mean bipartisan.

  15. Julie, since Baker v. Carr (1964 Supreme Court ruling), state legislators can’t design districts either for the US House or for state chambers that have substantially unequal populations. The inequalities referred to here are a result of population shifts since the last census.

    In any case, this problem–even were it deliberate–would be more accurately referred to as malapportionment rather than gerrymandering.

  16. Alan, I am not sure what the Shugart number is!

    Aaron, excellent points on the length of the reapportionment cycle, and the importance of it not being left to legislation to determine reapportionment. Some Latin American constitutions do not mandate a regular reapportionment, and hence have bad and increasing malapportionment even in their first (“popular”) chambers. France and Japan have also allowed malapportionment to get well out of hand before finally correcting it. (If I recall correctly, it was courts that forced the issue in Japan, while in France malapportionment was largely corrected after the Socialists won power in 1981.)

  17. I assume by “the Shugart number” Alan means the cube root of population. He didn’t like my “Round number” proposal (seats to equal the median between 200 and 1 per 100,000 – those numbers arbitraily chosen to suit the Aust H of Reps but easily adaptable to other polities).

  18. MSS, part of the issue with federal countries is that if they don’t allow districts to be in more than one state, then you can get significant inequities when rounding to get the number of districts. In the AU calculation I linked to, ACT is due 2.4 seats, and NT 1.5 – both get 2.
    In general, if you have more districts, then you are going to have more instances of unequal sizes due to unequal population changes (I assume). Therefore more seats should should be linked to shorter cycles.

  19. The Shugart number is mentioned on the MSS Wikipedia entry. (I do not think I put it there) My problem with the Round number is that while one can cite the empirical basis of the Shugart number the Round number has a large element of arbitrariness.

    As a sidebar, whenever, as they frequently do, politicians advocate smaller legislatures in the name of ‘small government’ they are actually reducing the number of their potential rivals.

    The small government mania does not seem to extend to ministerial or legislative stuffs. In Australia the ALP uses ministerial staffs as a feeder group for candidates and the numbers of staffers have expanded to create something very close to a full-blown nomenklatura in that party.

    It also seems important that redistribution be allowed on current figures, taking into account future trends. The UK practice of using start of review figures and then taking many years to complete a review contributes heavily to the level of malaportionment in that country. I suspect the US 10 year cycle has similar effects.

  20. ErrolC

    If you used a uniform quota across all states and territories, and alway rounded up, you would reduce the federal problem significantly. Of course that means the size of the legislature has to float with population rather than being a fixed number.

  21. Alan, by uniform quota you mean that the size of districts is determined by specifying the number of voters per district, rather than dividing a fixed number of seats by overall voters? So this would assist with the federal problem while population is rising.

  22. Yes, using a fixed number of seats inherently creates unequal quotas. The ACT and the NTA are apportioned seats under a rule that is formally fair and makes no distinctions between them. There is a large disparity between the ACT quota and the NT quota, because of the way rounding works. Even a rule that you always round up would significantly reduce disparity between quotas.The same principle operates to make Victoria’s quota slightly higher than NSW, although you’d expect the largest state too have the largest quota.

  23. Alan, I must be missing something, as I don’t see how uniform quota (separate from more representatives and rounding up) helps the federal problem. How are you setting the size of the uniform quota?

  24. Good. It was, after all, an early finding in a career of studying electoral systems and related institutions, which led up to his wining the Skytte Prize (2008).

    If I recall correctly, Rein’s first work on assembly sizes is from 1973. I was born in 1960, and while I am pretty sure I already knew in 1973 that the rules of elections made a difference, I am just as certain that I did not yet know what a cube root was.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s