Year of the Gerrymander

Michael Cain

Michael is a systems analyst, with a taste for obscure applied math. He's interested in energy supplies, the urban/rural divide, regional political differences in the US, and map-like things. Bicycling, and fencing (with swords, that is) act as stress relief.

Related Post Roulette

49 Responses

  1. Oscar Gordon says:

    Stephen Jay Gould & Niles Eldridge came up with Punctuated EquilibriumReport

  2. LeeEsq says:

    Roberts has been known to come to real howlers with electoral law and voting rights in a particular partisan manner. Shelby County is the foremost example. However, he does have a limit when the arguments made by his partisan side get too ridiculous like both healthcare cases or Trump’s massive argument on denaturalization. The fact that the Supreme Court, in the person of Alito, denied cert to in the Pennsylvania case is a sign that they seem to be growing tired of partisan gerrymandering and won’t tolerate it.

    Geographic based seat distribution will always hurt people who live in the denser populated areas. Since Democratic voters tend to cluster together for a variety of reasons, the Democratic Party will tend to get hurt. Where the Democratic Party does well on state and federal elections is where the state is geographically small and/or the Democratic Party is simply the biggest party. Maryland, New Jersey, Mass., California, are examples of this. What is really needed to make elections reflect the actual composition of the body politic is some type of proportional representation. That isn’t happening though.Report

  3. Pinky says:

    I don’t see how the Court can claim authority to determine the validity of districts using methods that aren’t in the Constitution, state constitutions, or law. Statistical measures provide a veneer of precision to what is a very subjective concept. If there were one goal for districts, we could compare a proposed map against that goal and see how well it matched. But districts may follow traditional lines, or lines of lower-level districts (such as a Congressional district representing three counties), or it may follow population centers, or geography, or demographics. It’s reasonable that a proposed map would be a compromise between a lot of different approaches, so why should we measure it against one specific statistical standard?Report

    • Michael Cain in reply to Pinky says:

      One of the conclusions of the district court in Wisconsin is that given today’s big data, software, and essentially unlimited processing power, it is possible to draw districts that meet any of the conventional measures — contiguous, compact, minimize splitting of counties and cities — and still have a partisan gerrymander. I doubt the Supremes are going to deny this.

      The amicus filing by Sam Wang et al. in the Wisconsin case includes three different statistical measures developed at the Princeton Election Consortium. Wisconsin fails on all three.Report

      • PD Shaw in reply to Michael Cain says:

        Wang filed alternative tests for the SCOTUS to use because he thought the notion of balancing “wasted votes” construct would be very unattractive to the SCOTUS. Thing is though, evidence on the Wang alternatives weren’t taken in the lower court. Will that matter?

        To me, the problem with these cases is that they are challenging a decennial redistricting near the end of the decade. Calculating wasted votes appears to be a post hoc test based upon observing voting outcomes across six years. I don’t think the courts will likely accept that type of statistical approach, and probably they shouldn’t. Electoral maps degrade, directing a remap can give the party in power the opportunity to mend their hold.Report

        • Michael Cain in reply to PD Shaw says:

          Not the same situation, but I’m reminded of the language when the Court approved the Cross-State Air Pollution Rule. Paraphrasing heavily: “We’ve ruled that there exists some stochastic model that meets muster. We’re not particularly happy with this one, but we can’t just keep sending this back to the lower court with ‘Sorry, not this one either, try again.’ So despite some misgivings, we’ve decided to go with this one.”Report

      • Kazzy in reply to Michael Cain says:

        Fivethirtyeight did a whole series on Gerrymandering and had an interactive map that looked at how things would shake out based on 7 or 8 different criteria for “optimizing” districts.Report

        • Marchmaine in reply to Kazzy says:

          The weird thing about that map… every scenario – except the Democratic Gerrymander – had congress going to the Republicans.

          I could easily be persuaded to some sort of mathematical framework… but which, exactly? On the one hand, “Highly Competitive” sounds kinda fair, but when I looked at my home state, where I have some familiarity of the geography and communities… the “Highly Competitive” model looked like we would have some very serious regional priority issues no matter who won which district… so in a strange way, that one didn’t pass my sniff test.

          The “Partisan Breakdown” model looked, well, about how it is now… with one battleground zone that one of the teams would not find favorable from an evolving demographic point of view… which lead me to @pd-shaw ‘s question of which snapshot of partisan breakdown do we use? Last Election? The last decennial? A snapshot of this year?

          But yeah… fun map. Makes you think maybe the Dems should go shhhh… wait till 2020.Report

          • Kazzy in reply to Marchmaine says:

            At the risk of being banned from Whole Foods again, I’m more interested in process than outcome. I’d have to look back at the different proposals, but when I harp on gerrymandering, it isn’t because it screws the Dems; it’s because it seems fundamentally unfair and anti-Democratic. If a fair(er) and (more) Democratic process still resulted in a GOP majority, so be it.Report

            • Marchmaine in reply to Kazzy says:

              Whole Foods sold out… none of the cool kids go there anymore.

              Yeah… ultimately 538 wasn’t emphatic in any direction:

              “If ending gerrymandering means creating maps that simultaneously enhance competition, don’t benefit either party, promote minority representation and keep cities, counties and communities whole, then it is impossible to end gerrymandering.”

              Maybe give each team one thing they can eliminate and see what happens? Which would you eliminate to end gerrymandering?Report

          • Stillwater in reply to Marchmaine says:

            I could easily be persuaded to some sort of mathematical framework… but which, exactly?

            Imperfectly, redrawn district maps should reflect state-wide voting habits of the electorate wrt national-level representation. I mean, that’s the goal, right? So any viable solution will require past data to constrain rather than be used to exploit partisan imbalances in representation.

            Colorado is a useful if under-discussed example of this: a marginally left leaning state has 5 R and 2 D representatives.Report

            • Marchmaine in reply to Stillwater says:

              “Imperfectly, redrawn district maps should reflect state-wide voting habits of the electorate wrt national-level representation.”

              For what time period? The previous 10 years? The next 10 years? Last Election? This Election? Perpetual today?Report

              • Stillwater in reply to Marchmaine says:

                States get to redraw every ten years, right? So the data upon which districts are redrawn includes the last ten years.

                Add: Congressional district drawing is backward looking, the politics of establishing a majority is forward looking.Report

              • Marchmaine in reply to Stillwater says:

                What does last 10-yrs of data mean? If an area was Blue 10-years ago but solidly red now are we counting it as Red or is there some historical weight to Blue voters who either aren’t there or are now Red voters?

                Just to be clear… I’m not so much “challenging” as asking… I usually get the impression that the models work for a snapshot and I’m not even sure what you do with an historical collection of snapshots… or why.

                So, every 10-years we take a picture of the electorate and use those models for the next 10-years? I’m ok with that, but I’d like it spelled out as part of the process. So each x10 election is a special election that sets the rules for the next 10 years… that’s what we have right now.Report

              • Stillwater in reply to Marchmaine says:

                See the comment below. Apart from aesthetic issues, the identified problem with current districting is that general voting sentiment isn’t reflected in congressional representation. And presumably the basis for believing this is a sequence of shapshots in time.

                I mean, maybe it isn’t a problem after all. But if it is, the basis for believing so is backward looking.Report

              • Stillwater in reply to Stillwater says:

                Btw, I say all that fully understanding that the meta-topic itself is incredibly political and therefore impossible to completely resolve. But if I’m understanding the complaint correctly, a necessary condition on a viable solution is that House seats better reflect state-level voting distributions. Anything which fails to do so isn’t a solution at all.Report

          • On the one hand, “Highly Competitive” sounds kinda fair

            Where voters (the people) have been asked recently, in the creation of redistricting commissions, they have accepted priority lists that put competitiveness at the bottom, if included at all. There’s probably a western bias to that, since most such have been done by initiative. Population patterns in most of the American West tend to specific communities of interest: the relatively few urban cores, the vast swathes of nearly empty rural, leaving most of the problem as how to chop up the suburbs.

            Utah is an interesting western case. It’s the only western state I’m aware of where gerrymandering is in use intentionally: the Republican legislature pretty openly admits that they chop up Salt Lake City and the inner suburbs so that they can’t elect a Dem to Congress. Polling says that Utah voters would prefer an independent redistricting commission by a 3:1 margin. They’d probably have done it, but Utah’s initiative process doesn’t extend to amendments, which is what would be required.Report

            • Marchmaine in reply to Michael Cain says:

              Yep, it was pretty clear in VA that you had to sacrifice “Communities of Interest” to get competition…

              But honestly, the idea of 50 “independent redistricting commissions” using 50 different methods doesn’t excite me. You can say, yeah, but that’s what we have right now, but without the independent part and you’d be right… but I’m not wild about that either – so I’m gettable for some sort of fair reform.

              I’d rather we have a model, an agreed upon timestamp, an agreed upon duration/reset, and off we go.

              Let’s have a good clean fight about the models and timestamps… and no commissions whose entire job is to re-litigate every demographic blip. But that’s just my preference… I have a high degree of doubt that the “independent commissions” will remain independent…Report

            • DavidTC in reply to Michael Cain says:

              Where voters (the people) have been asked recently, in the creation of redistricting commissions, they have accepted priority lists that put competitiveness at the bottom, if included at all.

              Count me as one of those. Competitiveness is not useful.

              While it might be hard to remember, the divide between the parties is completely arbitrary and artificial and saying ‘Every single district should be 50/50’ is nonsense. On top of that, thanks to incumbency, that isn’t how Congress works, so in reality all 50/50 districts would do is result in a randomly selected Congressentity from one party or the other, and then we keep that one for decades.

              There is no actual problem in creating a district of generally like-mind people, who will be pretty united on the sort of things they want in an elected official. In fact, this seems _much better_, representatively-speaking, not only because the elected official will do what the vast majority want, but also because that allows even finer dividing lines to be drawn…now candidates can quibble over exactly how to accomplish the community goals, instead of having huge fights over what those goals are.

              Population patterns in most of the American West tend to specific communities of interest: the relatively few urban cores, the vast swathes of nearly empty rural, leaving most of the problem as how to chop up the suburbs.

              I’m pretty certain that’s a mostly accurate way to describe the entire country. The West tends to have _emptier_ areas, or larger empty areas, and less cities, but in general that’s correct for the entire country.Report

              • PD Shaw in reply to DavidTC says:

                I guess my concern is that the “wasted votes” test is going to create more districts like the former rabbit-on-skateboard shaped district. The reason by such an odd shaped district is that downstate Illinois might be something like 60 to 40 Republican, but to get a reliable Democratic district they had to connect various constituencies: river rats, Swedes, African-American neighborhoods, coal country, and college campuses.

                My problem with the lack of compactness is that people become removed from their representative. Spreading the district across six different metropolitan statistical areas means at least six different media markets, most of whom will tend to ignore the politician from outside the area. But if the goal is to get maps that reflect the large number of downstate Democrats, funny shapes were likely needed. I just don’t think those types of districts elect good representatives and in the bunny-rabbit case it eventually elected a Republican anyway.Report

  4. Kazzy says:

    This past summer we went up to Marblehead, MA for our usual family vacation. While out with one of the boys, we stumbled upon a school named for the aforementioned Gerry character.Report

  5. Jaybird says:

    I said this back in 2011 and I stand by it:

    The solution to Gerrymandering is simple: Repeal the Apportionment Act of 1911 and return us to the level of representation we had 100 years ago (soon to be 101).

    Back in 2016, I got into it a bit more:

    My plan is just to eliminate both the Apportionment Act of 1911 and the Permanent Apportionment act of 1929.

    While I’d prefer to go back to 1789 levels of representation, I’d be down with saying that the least populous state in the union (Wyoming) has one representative and we round up for additional representatives. (So if Wyoming has 585,000 people, a theoretical state with 600,000 people would have two representatives.)

    And just give one electoral vote for each representative and each senator (*FOR*, not *TO*).

    Added bonus: eliminate gerrymandering.

    Additional added bonus: no need for Constitutional Amendment.

    My feelings on this haven’t evolved that much.

    I still think that we ought to go back to 1790 levels of representation.

    What is the current average number of constituents per representative? 709,760.

    What was the level of representation back in 1790? 34,436.

    We go back to one representative per 34,436 people, I guaran-damn-tee you that gerrymandering will evaporate.

    But I’d still settle for a Wyoming solution.Report

    • Jaybird in reply to Jaybird says:

      (This was hung up in moderation due to too many links. I think it will fix the problem though suffers from “never gonna happen”ism.)Report

    • Road Scholar in reply to Jaybird says:

      I’m on board with more Reps => lower constituent:representative ratio. I’m just not clear on how that alone solves (or materially improves) the gerrymandering problem. Is there evidence that states with either more total reps or a better ratio are less prone to this? If no evidence, what’s your logic chain?

      I believe a much more fruitful approach is the one put forward by a couple of congresscritters recently to create multi-member districts with each individual voter still casting a ballot for only one choice (optionally with an IRV/Ranked ballot mechanism).

      The specific proposal calls for districts of 3-5 members each. Basically take the existing number of reps for a state, divide by 4 and round normally. Any state with <=5 reps would have one big district. A state with 6 would have two districts with 3 reps each, etc.

      So, for example, Kansas with 4 reps total would be one big district. The four highest vote getters (preferably after IRV redistribution) would be seated. This would most likely result in a couple of Republicans and a Democrat with the last seat a R/D/L/I toss-up. Connecticut, again with four, would just be reversed.

      1. The House would be a lot closer to proportional albeit with a bit of an anti-majoritarian bias.

      2. Third parties – Libertarian, Greens, etc – would have a real shot at capturing some of those 3rd/4th/5th place slots.

      3. The two majors could very conceivably fracture into their natural components, giving voters a better selection toward their true leanings.

      4. Gerrymandering, if not impossible, would be much more difficult, tricky, and lower yield.

      5. Democrats in red states and vice versa may still be outnumbered in their state delegation but they would at least be represented. Better if we double or triple the number of reps so that the least populous states have 2 or 3.

      6. No constitutional amendment required for implementation.Report

      • Jaybird in reply to Road Scholar says:

        I’m just not clear on how that alone solves (or materially improves) the gerrymandering problem.

        I was thinking about someplace like Maryland. Maryland currently has 8 congressmen and their districts look like this.

        Maryland currently has 6.016 million people according to Wikipedia. (6,016,447. but who’s counting?)

        Going with the Wyoming solution (that is, the least populous state gets one congressman, everybody else gets one congressman, rounded up, for their number of this state’s populations… so, in this case, the population of Wyoming is 585,501 and a theoretical state with 600,000 people would have two congressmen) then we’d see that Maryland would have 10.27572, ah heck let’s round it up to 11 congressmen.

        That might help a *LITTLE* with the whole fractal districts thing, but my 1790 solution has us end up with 174.something congressmen which round up to 175 congressional districts.

        It’s my position that it is not really particularly possible to carve up 175 congressional districts in Maryland and maintain anything *CLOSE* to gerrymandering. I suppose it’s theoretically possible…carve it up into a bunch of strips a quarter mile wide and 3 miles long or something like that… but we’d have people who live next to each other be represented by someone that they voted for (and were a lot more likely to know personally) just off of little more than the limitations on being able to carve up the map into 175 different districts.Report

        • Michael Cain in reply to Jaybird says:

          Not just harder to do, but less attractive to the House members: “Just give me a solid suburban district, don’t be trying to stick me with a bit the urban core or a handful of old farmers.” In fact, I’d probably bet that the House would wind up being dominated by suburban members, who would discover that they had their own interests separate from the urban cores or rural areas. Maybe it’s fantasy on my part, but my time on the staff at the Colorado legislature convinced me that suburban Republicans and Democrats can come to agreements much more easily than urban Democrats and rural Republicans.Report

          • Jaybird in reply to Michael Cain says:

            Oh, I can *TOTALLY* see how House members would *HATE* this idea.Report

            • Michael Cain in reply to Jaybird says:

              In all seriousness, yes, for multiple reasons for sitting members. J_A, who hasn’t chimed in on this post, thinks we’re all so much alike that structural changes are silly. You regularly bring up “divorce or war”, where even the divorce is acrimonious. I think there’s a number of long-term trends that suggest a peaceful partition.Report

              • Jaybird in reply to Michael Cain says:

                Quite honestly, I think that greater representation would provide one hell of a release valve.

                Imagine knowing your representative. Imagine him remembering you.

                Only the most privileged in our society can do that now. Those who are tied to PACs and SuperPACs.Report

              • J_A in reply to Michael Cain says:


                I read the post when it had zero comments, and thought about chiming in, but I would have gone so far away from the OP that I decided better.

                In a nutshell I believe the states (except for HI, AK, LA, eventually PR and the VI, and perhaps exceptionally UT) are anachronisms that made sense in the late XVIII century but stopped soon thereafter. There is no real difference between Maine and New Hampshire, and there’s really no substantial difference between OR and SC that requires SC to pass a bunch of law school that are substantially the same as the laws OR pass.

                In deference to @michael-cain, I fully understand that a handful of matters, mostly related to the actual geographical environment, are different in the West than in the Atlantic Coast, and should be addressed differently. Yet, even so, I’m not sure that many of these issues encountered in NV are not similar to those encountered in KS, for instance. (*).

                But the rest of what we do is exactly the same. I’ve lived in two states, and many of us have lived in many more, and becoming a TX citizen required me packing my stuff from FL and getting a new drivers license. My opinions about how many years for burglary, accounting rules, or the distance between buildings didn’t change when I became part of a different state.

                Tl/Dr. The differences within states are just a tool to gain electoral power (“ We are SC, we don’t let NC tell us how to let people vote”

                The only real difference is urban, suburban, rural

                (*) Disclosure, I know upstate NV all the way north to Ely, having deal5 with a power plant there, but I have never been to KS.Report

              • J_A in reply to J_A says:

                The only real difference is urban, suburban, rural

                And that’s how I would apportion the house(s)

                By creating compact districts of homogeneous population density. Those would represent people that really face the same issues day by day. This is what the Fathers had in mind when they said the House represented the People, while the Senate represented the Sovereigns (paraphrasing).

                Gerrymandering is based in representing the voters, and not the people. That’s where, IMANHATO, the problem starts. Republican or Democrat, representatives that vote against building highways in a compact suburban district won’t last long (not that I believe highways are good, just an example, guys). Likewise, urban representatives will vote for more funds for education and less funds for rural broadband. And that’s the way it should be. Rural representatives will fight for milk subsidies,

                Yes, the rural districts would probably be not very compact because they would grab all that is not urban or suburban, but in 0rincip,e their issues should also be similar.Report

    • Marchmaine in reply to Jaybird says:

      So at 34,436 people per district apportioned over ~330M people we’re looking at 9,582 congressmen and 100 Senators.

      That’s no longer a deliberative body, but maybe that’s a fiction we give up to readjust to mass society.

      I could see changing the nature of congress so that representatives don’t really leave their districts at all – all virtual… maybe a couple of bi-annual conventions or some such. Hard to fathom the types of new games we’d see, but part of me is perfectly happy to experiment; the other part is perfectly happy to experiment in, oh I don’t know, Canada.

      I’d also see this potentially leading to more decentralized regional decision making… breaking down the legislature into regional working groups of 1,500-2,500 members. All sorts of new permutations…

      p.s. Hard to believe the senate doesn’t get nuked… imagine having to get ~5,000 yeses just to have 40 Senators stall an agenda? Yeah, that has nuke bait written all over it.Report

      • Jaybird in reply to Marchmaine says:

        p.s. Hard to believe the senate doesn’t get nuked… imagine having to get ~5,000 yeses just to have 40 Senators stall an agenda? Yeah, that has nuke bait written all over it.

        If they can hold their attention that long.

        For the record, I’d only want to do this if we go back to the old way of doing the Senate.

        Ooooh! Something else that’s fun to consider:


        • Road Scholar in reply to Jaybird says:

          It reduces the small-state advantage.Report

        • Marchmaine in reply to Jaybird says:

          Wait, we still have only one president?Report

        • Jesse in reply to Jaybird says:

          The old way? The part where gerrymandered state legislatures hand the Senate over to whomever gives them the most money?

          I still don’t get the libertarian or in Jaybird’s case, former libertarian’s obsession with state legislators whom nobody knows or likes deciding who the Senator is.

          Hell, state legislatures didn’t even like it, since state legislature races became about who would appoint the Senator, not say, actual state level issues.

          Also, as always, the Senate must be destroyed.Report

          • Jaybird in reply to Jesse says:

            It’s an idea that works hand in hand with the whole “change the makeup back to the way the Founders had it at the beginning” thing.

            Go back to the 1790 way of doing the House.
            Go back to the 1790 way of doing the Senate.

            I also have the comments I gave in response to you saying “I don’t get the arguments that people who aren’t me have!” last time, if you want me to google them and link you to them.Report

            • j r in reply to Jaybird says:

              Why would we want to go back? If anything, we should want to take governance forward.

              Why aren’t more things happening online? Why does so much of the business of government have to happen in a a few big neoclassical and brutalist buildings scattered around a big plot of grass in Washington?

              You can preserve the principles of representative, deliberative democracy in a form that is workable in a modern context. Heck, you may even be able to make improvements. But I don’t see how turning DC into the Galactic Senate would be an improvement; it certainly didn’t improve the Star Wars prequels.Report

              • Jaybird in reply to j r says:

                I think we’ll have better luck repeating the mistakes of the past than we will making new mistakes that we’ve never made before.

                As for the online thing, I think that the online thing could work well with having more representatives. But the senate should still be deliberative in the big brutalist buildings.

                The prequels had many problems and the Senate was the least of them.Report

              • j r in reply to Jaybird says:


                I think we’ll have better luck repeating the mistakes of the past than we will making new mistakes that we’ve never made before.

                I honestly don’t know what this means.Report

              • Jaybird in reply to j r says:

                Well, there are two options:

                1. Keep doing what we’re doing
                2. Change what we’re doing

                If we do the second, there are two ways to do that:

                1. Do something that we did before
                2. Do something we haven’t done before

                I’m thinking that we shouldn’t keep doing what we’re doing and that we should change… but we should go back to what we had before rather than, I guess, blowing up the senate and having courts choose the districts based on some weird algorithm that only coincidentally helps their preferred politicians?

                I’d rather go back to representatives who represent a number of constituents in the mid-30Ks (rather than more than half a mil!) and a deliberative body chosen by professional politicians because I think that that will be a *LOT* harder to game than most anything else I’ve seen other people recommend.

                Certainly harder to game than what we have now, I tell you what.Report

            • Jesse in reply to Jaybird says:

              Or ya’ know, we can look back and return to the old ways when they make sense and not go back to the old ways when they were incredibly stupid.

              Should we also bring back the 3/5 Compromise since that was how they did it in 1790 as well?Report

          • Michael Cain in reply to Jesse says:

            So, clearly you disagree with what I think the Supremes are going to say later this term about gerrymandered state legislative districts. As Amarillo Slim said, bet or shut up :^) What’s the split, who’s on which side, who writes the opinion?Report

          • Mike Schilling in reply to Jesse says:

            The old way? The part where gerrymandered state legislatures hand the Senate over to whomever gives them the most money?

            You say that as if it weren’t the whole point.Report

  6. Pat says:

    Change the House to proportional party representation. Gerrymandering disappears: no more geographical districts for the House.

    You cast your vote for a party, like in a parliamentary system. The parties pick their representatives. Not as much room for ideological variance within a party, but the minor parties will (a) get meaningful representation (b) have to do something with it rather than caper on about how they’re the Real Progressive/Conservatives and (c) will have an easier time accepting compromises from their representatives because they won’t have the RINO/DINO accusation to fall back on.

    Jaybird’s solution is fine with me, too, and I find the objection “but then Congress will be huge” not particularly problematic. The subcommittee structure and caucuses will do their thing.

    But it will be much harder to maintain monolithic party control, and much more room for ideological variance.

    What we have currently are huge structural incentives to have only two parties, and very weak party controls and party minorities forcing ideological purity through structural shenanigans (Hastert Rule, fillibuster, gerrymandering, etc).

    So you either need to remove the structural rules that enforce two parties (proportional representation) and enable multiple parties who can negotiation (meaningfully) inter-party differences, or you need to increase representation such that it’s just practically very difficult to use structural shenanigans.Report