The Power of Gerrymandering

With demographic realities making Republican wins at the polls less and less likely, their electoral strategy has been a two-pronged one: 1) voter suppression tactics to keep as many Democratic voters from the polls as possible, and 2) gerrymandering congressional districts to isolate Democratic voters and create safe Republican seats. The first part didn’t work out all that well in 2012, but the gerrymandering has been a huge success, as Sam Wang of the Princeton Election Consortium documents.

Through artful drawing of district boundaries, it is possible to put large groups of voters on the losing side of every election. The Republican State Leadership Committee, a Washington-based political group dedicated to electing state officeholders, recently issued a progress report on Redmap, its multiyear plan to influence redistricting. The $30 million strategy consists of two steps for tilting the playing field: take over state legislatures before the decennial Census, then redraw state and Congressional districts to lock in partisan advantages. The plan was highly successful…

We can calculate each state’s appropriate seat breakdown — in other words, how a Congressional delegation would be constituted if its districts were not contorted to protect a political party or an incumbent. We do this by randomly picking combinations of districts from around the United States that add up to the same statewide vote total. Like a fantasy baseball team, a delegation put together this way is not constrained by the limits of geography. On a computer, it is possible to create millions of such unbiased delegations in short order. In this way, we can ask what would happen if a state had districts that were typical of the rest of the nation.

In North Carolina, where the two-party House vote was 51 percent Democratic, 49 percent Republican, the average simulated delegation was seven Democrats and six Republicans. The actual outcome? Four Democrats, nine Republicans — a split that occurred in less than 1 percent of simulations. If districts were drawn fairly, this lopsided discrepancy would hardly ever occur…

Gerrymandering is not hard. The core technique is to jam voters likely to favor your opponents into a few throwaway districts where the other side will win lopsided victories, a strategy known as “packing.” Arrange other boundaries to win close victories, “cracking” opposition groups into many districts. Professionals use proprietary software to draw districts, but free software like Dave’s Redistricting App lets you do it from your couch.

So how well does this strategy work to make the outcome of congressional elections not fit with the overall popular vote? Incredibly well.

We can quantify this effect using three different methods. First, Democrats would have had to win the popular vote by 7 percentage points to take control of the House the way that districts are now (assuming that votes shifted by a similar percentage across all districts). That’s an 8-point increase over what they would have had to do in 2010, and a margin that happens in only about one-third of Congressional elections.

Second, if we replace the eight partisan gerrymanders with the mock delegations from my simulations, this would lead to a seat count of 215 Democrats, 220 Republicans, give or take a few.

Third, gerrymandering is a major form of disenfranchisement. In the seven states where Republicans redrew the districts, 16.7 million votes were cast for Republicans and 16.4 million votes were cast for Democrats. This elected 73 Republicans and 34 Democrats. Given the average percentage of the vote it takes to elect representatives elsewhere in the country, that combination would normally require only 14.7 million Democratic votes. Or put another way, 1.7 million votes (16.4 minus 14.7) were effectively packed into Democratic districts and wasted.

Yes, Democrats have done it too, but not nearly as often and not nearly as effectively. This is why the drawing of congressional districts should be out of the hands of elected officials and political parties.


  1. raven says

    Yes, Democrats have done it too, but not nearly as often and not nearly as effectively.

    If the GOP cheats (legally), it all but forces the Democrats to do the same.

    Given the impending death of their base, old white male christofascists, it will happen sooner or later.

    Live by the Gerrymander, die by the Gerrymander.

  2. says

    Can anyone point us to any particular legislative proposal that describes an effective alternative method of drawing representative district lines? It’s one thing to describe the problem — but the Devil is in the details when it comes to legislating a solution.

  3. abb3w says

    Yes, Illinois is one of the states massively Gerrymandered. However, it was actually at the low end of the nine cases he highlighted; the GOP Gerrymandered more states and in larger degree than the Democrats.

  4. brucegee1962 says

    Can anyone point to a place where a law is proposed that would prevent gerrymandering? The idea that it should be done by “non-partisan” groups seems unworkable — there aren’t any such things in America. The only thing I know of is a strictly geometrical test — something like “The distance from the center to any border of a district may not be more than 75% or less than 25% of the average distance.” That seems clunky, but maybe workable. Does anyone know of a better way?

  5. abb3w says

    …though, lest mere mention be confused for advocacy, I feel compelled to note that Mencken’s observation “For every complex problem there is an answer that is clear, simple, and wrong” gives me considerable pause about the Shortest Splitline approach.

    I’d recommend anyone with serious interest look through Sam Wang’s blog, as well as the NYTimes piece.

  6. baal says

    Court cases on redistricting usually have sets of alternative maps drawn up. It’s not all that hard to then select among the maps based on one criteria or another. It’s an easy empirical question to select a map based on which map give you district seats closest to the popular vote split (i.e. least gerrymandered). The link in abb3w’s post talks more about this approach. There is some math involved but it’s not actually all that hard a problem if you wanted non-gerrymandered maps to figure out what they’d look like.

  7. laurentweppe says

    The [voter supression campain] didn’t work out all that well in 2012
    Democrats would have had to win the popular vote by 7 percentage points to take control of the House

    Of course, if all the minorities who lost the right to vote because they were sent to prison after a peccadillo for which no middle-class & above white who ever be punished or after a cop testylied against them, a 7 point advantage would not be so far-fetched. i’d say that voter suppression is working quite well


    Illinois is one of the states massively Gerrymandered

    Yes, and the Democrats gerrymandered Chicago, you know, the place they dominate so much that they decided to force to mayoral candidates to run as non-partisan in order to hide the fact that all candidates are democrats. While the Republican cheat to preserve their power and influence, Democrats rig elections they’re sure to win fairly: how mad is that?

  8. Jordan Genso says

    Raging Bee

    Can anyone point us to any particular legislative proposal that describes an effective alternative method of drawing representative district lines? It’s one thing to describe the problem — but the Devil is in the details when it comes to legislating a solution.

    The other problem with legislating a solution is that the legislation doesn’t have to be followed. As a member of the five person panel that reapportioned my county in 2011, I was naive enough to think the other four (the chair of the county GOP, the county clerk, treasurer and prosecutor – all Republicans) would follow the law. Little did I know one important thing- if they break the law, there is a huge cost in trying to take them to court, and while the county tax dollars would be used in defending their criminal action, my group (the county Democratic Party) would have to pay all of the costs for our side.

    And so, we get a very obvious and clear example of how gerrmandering happens on the local level:

    I’m the bearded one in the middle. The legislation was very obvious in how we had to act as a board, by following the guidelines in the legislation, but they didn’t seem to care.

  9. vmanis1 says

    I don’t think there’s a single thing that can eliminate all gerrymandering, but there are a group of steps that can between them reduce it.

    1. Make the preparation of proposed electoral maps the responsibility of a (non-partisan) election agency, under the supervision of a non-partisan board (retired judges work well in jurisdictions where judges aren’t elected).

    2. Use an open algorithm (not necessarily Shortest Splitline) and clearly-defined criteria (e.g., population density, geographical remoteness) to define electoral boundaries. This way independent individuals (e.g., political scientists, campaign directors in the political parties) can more intelligently critique any proposals.

    3. When a proposed map is produced, the legislature can only accept or reject it, not amend it.

    The goal is to get as much partisan control out of the process as possible. Not perfect, but maybe a bit better.

  10. Childermass says

    One would think that one could write an algorithm that could be used in any state that is independent of anyone’s political wishes. Just input the data of where the citizens live, the shape of the state, city limits, county lines, locations of interstate highways, and rivers, etc. and have the program spit out the results. Where individual people live at the time of the census should be the only input relating to humans i.e. no data of age, race, sex, occupation, what party someone signed up for, etc. can be included.

    As in theory, political boundaries could be monkeyed with to subvert the program, changes in political boundaries and geographic boundaries will be inputted after a few years have passed. Also program should randomly decide which of a set of possible weights to the factors used by the algorithm. That would require that running of the program be done in public with representatives of any political party who won at least 5% of the vote in the last federal election having a guaranteed spot at the proceedings.

    Finally, we really should push for the House of Representative to be increased to 1000 members. It would be much harder to gerrymander that and would decrease the amount of money needed to run for Congress. Obviously the full increase should not be done at once, but rather in stages so the fortunes of the parties in one particular election are not frozen in place.

  11. says

    Childermass: if such algorithms are used, then the specific rules, and the source-code for any computer programs used, would have to be published for all to see, along with verification that those are the steps actually followed, and all of the specific steps leading to each change of district boundaries. There would probably have to be some kid of protest/appeal process — but in that case, remedies should be restricted to a complete reiteration of the procedure, and not allow for piecemeal fiddling with lines here and there as part of a political settlement or deal.

  12. Rip Steakface says

    The idea that it should be done by “non-partisan” groups seems unworkable — there aren’t any such things in America.

    Washington state does its districts that way. Refer to the WA Redistricting Commission.

  13. Michael Heath says

    abbeycadabra writes:

    Why have districts at all? Why not elect all the representatives statewide on one multiple-position ballot?

    Because members of Congress service a specific group of people. Members of Congress don’t merely legislate, they also have offices throughout their respective district that service the requests of their constituents.

  14. Mark Sherry says

    If the redistricting simulations described in the article give fairly consistent results, you could have a requirement that if the previous election were run using the proposed redistricting, it cannot deviate in results from the simulated models by some amount. There’s still room for playing games, but there’s a bound on how much of an advantage either party is likely to gain, at least without taking future demographic changes into account.

    The amount of partisan control in the US electoral process astounds me. Perfect neutrality is impossible to achieve, but the current system seems designed to cause the current abuse we’re seeing. It’s like a textbook case of “How not to handle redistricting”, and since both parties look forward to abusing it when they’re in power, there’s no political will to fix it. (Plus even if the Democrats did decide to try fixing it while in power, there would be great cries of protest from the other side. “Keep the government out of our elections!” “I want the good God-fearing man I elected to determine where the districts should be, not some faceless Government bureaucrat!” “The Federal Electoral Commission is a front for the Muslim Brotherhood/ACORN/The Teamster’s Union/Planned Parenthood/The Communist Party of America (cross off those which do not apply)”)

  15. lpetrich says

    Multimember districts would be a good way to improve representation, with appropriate voting systems.
    Single Transferable Vote is like Instant Runoff Voting, but with winners as well as losers eliminated as one goes, and some of the winners’ votes reused for other candidates. It’s used in Ireland and Australia.
    Party List proportional representation is widely used, and mixtures of it and single-member districts also have a lot of use.

  16. whistlepete says

    “This is why the drawing of congressional districts should be out of the hands of elected officials and political parties.”

    While I understand the need for that I don’t think that it is possible. Almost everyone has political leanings to one side or the other. I don’t see how this could be done apolitically. Each side is always going to try to gain an advantage over the other side. Maybe democrats will just have to get better at it or figure out a way to lesson the effects of it.

Leave a Reply