Fighting Gerrymandering with Mathematics
February 11, 2017 1:31 PM   Subscribe

A 5-day summer school will be offered at Tufts University from August 7-11, 2017, with the principal purpose of training mathematicians to be expert witnesses for court cases on redistricting and gerrymandering. How gerrymandered is your congressional district, anyway? You can use geometry as a proxy: compare the perimeter of your district to the perimeter of a circle with the same area. The Washington Post suggests what non-gerrymandered districts might look like. Previouslies, especially compactness
posted by leahwrenn (75 comments total) 70 users marked this as a favorite

 
Dumb question, but...why aren't counties the same as districts? For rural counties, they can be added until a threshold level, and urban counties can be split. Wouldn't that be simplest?
posted by zardoz at 1:43 PM on February 11


zardoz, how you choose to add (or subdivide) counties leads to exactly the same problem. Maryland is gerrymandered to hell and back. It has 24 counties ranging in size from 20,000 people to just over one million. From that you need to extract eight US representatives.

Good luck.
posted by It's Never Lurgi at 1:52 PM on February 11 [7 favorites]


You *can't* use geometry as a proxy. (This point repeatedly comes up in the "compactness" previously link).

Democrats cluster more heavily than Republicans (at this moment in time). So geometrically pleasing districts won't necessarily accord with "fair" districts, where "fair" here means that the ideological diversity of the resulting elected body matches that of the constituents.

Arizona, for example, has its redistricting done by a non-legislative commission, and not so long ago won a Supreme Court case about it. As a consequence, its national house reps more closely align with its actual population than many other states with more smooth-looking districts (Michigan, for example).

From one of the previouslies, take this quiz and see if you can pick the gerrymandered map.
posted by nat at 1:58 PM on February 11 [25 favorites]


Gerrymandering and the right to vote are going to be two of the most important structural political fights in the next two years. Back in October it was reported that Obama and Holder were going to take up a redistricting effort, but I haven't heard any updates.

The compare-to-a-circle metric isn't so great; districts should not be featureless blobs. It makes perfect sense to draw a district boundary around a city, say, to separate their legislative concerns vs. surrounding rural areas.

A much better measure of gerrymandering is the efficiency gap, roughly a count of the number of votes suppressed by the districting plan and which party benefits. There's an effort underway to use this kind of measure in courts. More technical details in this paper.

Iit's important to have some historical context on gerrymandering. Back in the 70s the Democrats were the ones who used it more effectively to stay in power. Also redistricting in order to ensure minority representation in Congress has historically been a tool of progressivism.
posted by Nelson at 2:04 PM on February 11 [18 favorites]


How would this change the makeup of the House, I wonder? One of the criticisms of the current system is that the ratio of Democratic to Republican representatives nationwide doesn't reflect the popular vote for representatives. In 2016, I believe, the GOP got 48.7% of the "representative popular vote", to the Democrats 47.9%, but got 55% of the seats.

I'm not sure that inexact representation is, ipso facto, a problem (because if our goal is proportional representation then let's just freaking do proportional representation and forget about districts), but if it is then I think it's reasonable to ask how the new scheme stacks up. I haven't seen that analysis.

Maryland, to use an egregious example, has seven Democrats and one Republican in the House. It's a heavily Democratic state, but a proportional breakdown would be 6-2 or 5-3. What would the compactness version give us?

Edit: I went to Brian Olson's website and he specifically asks this question and says he doesn't know the answer and doesn't really want to (he doesn't want this to be about who will win and who will lose). That's fine, but if it turns out that these new districts resulted in a massive imbalance in either direction then that would be important information, right?
posted by It's Never Lurgi at 2:10 PM on February 11 [4 favorites]


Back in October it was reported that Obama and Holder were going to take up a redistricting effort, but I haven't heard any updates.

Here you go.
posted by escabeche at 2:47 PM on February 11 [5 favorites]


Iit's important to have some historical context on gerrymandering. Back in the 70s the Democrats were the ones who used it more effectively to stay in power. Also redistricting in order to ensure minority representation in Congress has historically been a tool of progressivism.

And it still has to end. It's not any more democratic when we do it than when they do, and it results, when the Republicans finally get power, in an attitude that they have to do it to make up for when Democrats did it.

Ultimately, these kinds of plans that thwart the popular will through gaming the system produce things (to pick a broadly if not precisely true example) like a newly-elected President with approval rates below 50%. Democracy isn't, and has never been, a magical way to give people good government, only that government they, in the aggregate, deserve.
posted by JHarris at 3:04 PM on February 11 [12 favorites]


only that government they, in the aggregate, deserve.

I used to agree, but the truth is, the American people have been working their asses off to improve things, but the political leadership has been actively working against them assuming themselves to know better. For example, I cite this one a lot, but it's an important case study: in Florida, we have a constitutional requirement to provide a well funded, universally accessible, uniformly high quality public school system. Years back, we were promised a boon of new education funding if we voted to create a state lottery. Around the same time, a lot of legitimate academic education researchers began releasing studies that showed having significantly smaller class sizes significantly improved education outcomes. Class sizes in Florida had been creeping up from historical norms because our political leaders were already slowly squeezing funds out of the system for other priorities or refusing to increase funding consistently year over year to account for population growth and inflation. So the people of Florida organized and overcame massive organizational challenges and deliberate procedural roadblocks to get an amendment to the constitution passed to mandate a specific smaller class size ratio (12 to 1, I think, or similar). The measure passed and became the strongest most unambiguous law of the land in the state of Florida it is possible to be: a constitutional requirement. Voters expected the Republican leadership in the state to follow through on their promises and supplement the existing education budget with the new lottery revenues to make the class size amendment financially workable, but instead, the new lottery funds were offset with new budget cuts, and the Florida legislature insisted it would be impossible to fund the new requirement and started using procedural tricks and judiciary processes to stall implementation. It's been over a dozen years since our democratic processes made smaller class sizes in public schools the law in Florida, but our legislature has refused to follow the law for over a decade now and looks to be close to permanently killing even any pretense of making compliance with the state constitution a priority.

The American people have still been doing their part. It's our political leaders who have lost respect for democracy and the rule of law.
posted by saulgoodman at 3:26 PM on February 11 [61 favorites]


Thanks for this! Here is another nice analysis, as well.

The issue of existing state borders doesn't seem too well addressed by a lot of these metrics. In the third link, the little methodology note mentions coastlines, but, e.g., WV-01's visually imperfect-but-reasonable boundaries still get their compactness score ruined by the little spiky bit at the top. Which is kind of a shame, since its score is similar to the (avoidable) nonsense fractal nightmare happening just east of it in Maryland.

A method that could ignore difficult borders would be neat - maybe something that preserves area but deforms state borders into friendly arcs for the purposes of calculating the compactness.
posted by nvvd at 4:23 PM on February 11 [1 favorite]


My district is made up of 8 pieces that don't even touch. Not trying to make a point or anything, but that's weird.
posted by FirstMateKate at 4:50 PM on February 11 [2 favorites]


Screw districts. Get rid of them altogether.
posted by kyrademon at 4:59 PM on February 11 [3 favorites]


I'm in District 1 in NC. 96.01 gerrymandering score (on scale of 100). That's one of the many reasons why 80,000 of us marched in Raleigh today! Sick of the GOP cheating the system when they can't win the votes.
posted by duvatney at 5:19 PM on February 11 [13 favorites]


I hope there are geographers involved. It seems odd to target this at mathematicians when there are lots of quantitatively and spatially trained geographers around who probably have a reasonably deep understanding of the socioeconomic, political, and racial aspects of redistricting.
posted by mollweide at 5:54 PM on February 11 [15 favorites]


Jowei Chen at UMich has been doing some interesting work on simulated redistricting. His paper Cutting Through The Thicket shows how you can arrive at different outcomes even when taking some basic principles into account such as keeping counties intact, not cutting through cities, and not breaking up districts mandated by the Voting Rights Act. Compactness and other aesthetic qualities of a district are terrible measures. For example, Florida’s District 5 looks like a serpentine mess but it’s required to be that way due to the VRA.

Efficiency Gap is one of the most promising measures right now because it uses actual election results and the relatable concept of a wasted vote to work. It’s very easy to calculate for a past election — you just compare votes that didn’t contribute to a win, whether they were cast for a losing candidate or in excess of the threshold for a winning candidate. It’s tricky to calculate for uncontested elections and proposed plans. There’s a bunch of imputation you need to do to come up with the right values.

I only know these things because I sent fan mail to Eric McGhee and he graciously helped me understand a bunch of the basic concepts over coffee.
posted by migurski at 6:05 PM on February 11 [8 favorites]


I would donate to this, or some effort to create scholarships or recruitment/identify-mathematicians-who-communicate-well efforts.
posted by amtho at 6:20 PM on February 11 [2 favorites]


What if we just did away with congressional districts and sent the top X vote getters to Congress? Wouldn't that weight every individual's vote equally within their state? Seems like it would also produce a more diverse group of representatives than we have now. I mean, Illinois has 18 districts. What sort of person would come in 18th? I bet they wouldn't be like anyone currently in Congress.
posted by great_radio at 6:23 PM on February 11 [4 favorites]


What if we just did away with congressional districts and sent the top X vote getters to Congress? Wouldn't that weight every individual's vote equally within their state? Seems like it would also produce a more diverse group of representatives than we have now. I mean, Illinois has 18 districts. What sort of person would come in 18th?

Probably somebody from Chicago, just like the other 17 top X vote getters... Downstate's got enough problems with being ignored by Chicago as it is.
posted by leahwrenn at 6:31 PM on February 11 [8 favorites]


I hope there are geographers involved. It seems odd to target this at mathematicians when there are lots of quantitatively and spatially trained geographers around who probably have a reasonably deep understanding of the socioeconomic, political, and racial aspects of redistricting.

So much this. The map generated by compactness algorithm feels all wrong, and I couldn't put my finger on it at first. But it's basically just a case of Engineer's Disease (even if it's coming from a mathematician).

Screw districts. Get rid of them altogether.

I don't entirely disagree, but I think there's something to be said for ensuring local representation at the federal level. To that end, we need more representatives. And as long as we're dreaming, I also favor the electoral system used for the German Bundestag.
posted by HumuloneRanger at 6:36 PM on February 11 [4 favorites]


“Let’s not have districts” seems like a non-starter to me. For one, the strategies advocated by Indivisible only work when direct constituents of a representative show up to talk to them. Republicans have cheated at this game, most directly in the 2010 redistricting with REDMAP. It won’t work to change up the rules to make cheating impossible; there’ll just be new ways to cheat. Instead it must be possible to identify and fix cheating.
posted by migurski at 6:40 PM on February 11 [4 favorites]


The districting process itself not only tends to lead to these shenanigans but also enshrines the two existing parties. Getting rid of districts all together and running at large would eliminate the motivation for shenanigans. On the other hand, as others noted, it would tend to favour more populous areas over less populous. It should also make it more difficult for independent candidates to win any seats (having to battle at a larger scale, and eliminating the chances of a candidate with strong regional popularity). A parliamentary system where representatives were allotted according to their party's proportion of the national vote might address some of the failings of the current system, but remove representatives from accountability to a particular constituency and concentrate their efforts on the most populous areas.

Solution: Enforced redistribution of the population until any subsection is within epsilon of the average population density for the whole.
posted by oheso at 7:05 PM on February 11 [1 favorite]


Jeff Sessions was obviously the bagman for voter supression and federal gerrymandering. He will generate a voter fraud scandal in 6 months or so (well, after a new SCOTUS pick) to allow federal voter suppression. The opposition needs to be ready for this shit.
posted by benzenedream at 7:18 PM on February 11 [3 favorites]


I live in district 1 of Vermont, and we're not gerrymandered at all.
(Just happy to be represented at all, compared to when I lived in DC)
posted by MtDewd at 7:21 PM on February 11


"Dumb question, but...why aren't counties the same as districts? For rural counties, they can be added until a threshold level, and urban counties can be split. Wouldn't that be simplest?"

In Illinois, Cook County has 5.2 million people. Hardin County has 4,135. That's the essential problem. And regions may share different things -- Peoria shares a broadcast market with Bloomington/Normal to the east, but it's more industrially similar to the Quad Cities (NW) or Decatur (SE). Its transit concerns run up and down the Illinois River (NE and SW) with still a third set of counties, but its farming more closely matches counties to the west.

Illinois has 18 congressional districts, 12 in the Chicago area and 6 to cover the entire rest of the state, with 96 non-Chicago-area counties, 8 cities over 100,000, and a further 31 over 50,000 (but less than 100,000).

My congressional district covers 6,933 square miles, and it's fairly compact for the six downstate districts. (The others cover 14,696; 7,918; 10,516; 5,008 (east st louis); and 5,794 (champaign-urbana) square miles.)
posted by Eyebrows McGee at 7:35 PM on February 11 [4 favorites]


"In Illinois, Cook County has 5.2 million people. Hardin County has 4,135. "

And I should be clear -- Illinois is relatively densely populated (for being west of the Appalachians!). If you get out into the real West, Colorado's smallest county is 699 and its largest is 622,263; Hawaii's smallest is NINETY people, while its largest is 953,207; Texas's smallest is 82, while its largest is 4,092,459!

New York has a startling 69,468 people per square mile in Manhattan (NY County), and just 2.81 per square mile in Hamilton County!
posted by Eyebrows McGee at 8:09 PM on February 11 [4 favorites]


Just to pile on:

Loving County, Texas: Population 82
Harris County, Texas: Population 4,538,028

I'm starting to fall into the camp of eliminating geographic districts completely because the temptation to abuse the system is too great. My proposed alternative is that all registered voters in the state get a representative assigned to them at random. That seat becomes, in effect, your "district." New voters are also assigned at random, perhaps with a small bias to keep the numbers trending towards even.
posted by LastOfHisKind at 8:11 PM on February 11 [5 favorites]


On the more main topic of the post -- When I was on school board I was actually involved in a decennial redistricting, which in our case (which is unique in the US in its requirements) had to meet certain requirements set by the federal courts in a federal consent decree dating to the Civil Rights era, and then had to meet the typical rules of compactness and contiguousness and equal population.

To manage it, we hired a specialist mapping company that does basically nothing but governmental districts (voting, schools, etc.), and then a specialist attorney to review it who specializes in Voting Rights Act compliance. It was interesting because with computers it's a LOT easier than it was 20 years ago, but in the end there is still an actual human familiar with the area looking at streets and houses by hand and making sure the boundaries make rational sense, house by house by house. (It's the same process for school boundaries, but worse, because families cluster so it's very hard to equalize school populations (some areas have tons of kids and some hardly any) and simultaneously make walking/bus routes not suck!)

It was really interesting, how difficult it was to draw three districts (for seven representatives -- 3 and 3 and 1, it's a weird system) in a city of 110,000 that met all the federal requirements and made "sense" in terms of city geography. It was a glimpse into how difficult it must be when you're talking about a whole state, and millions of people, and many more subdivisions.

A couple years later I got jumped over when my congressional district line moved four blocks, moving me from the GOP gerrymander to the Dem gerrymander, which was very pleasing. But basically the same process, computers and specialists and someone checking it house by house by house.
posted by Eyebrows McGee at 8:19 PM on February 11 [5 favorites]


Tufts has been a leader in GIS Data Rescue too, thanks for this heads up OP.
posted by drowsy at 8:25 PM on February 11 [3 favorites]


It's not a good idea to eliminate legislative districts - because geography matters. There are issues involving land use, population density, watersheds, resource extraction, agriculture, transportation corridors and any number of geographical and environments characteristics that affect people and their lives and which legislators must be familiar with. In most states, people in rural areas would have no representation, and the policies set would only be appropriate for urban and suburban areas. What agricultural policies would be set in a state where none of the legislators lived in agricultural areas?

We shouldn't give up on trying to draw districts that are more fair just because there is no perfect method of doing so.
posted by tommyD at 8:55 PM on February 11 [6 favorites]


Part of the problem is the size of the districts. Since we limited the number of representatives at 435, the districts are now so large that it is practically impossible to create a district that has any sort of cohesion on geographic, political, racial, or any other representative basis.

The more I think about the problems facing the country, the more I think a constitutional convention is in order.
posted by Big Al 8000 at 10:24 PM on February 11 [3 favorites]


Came into the comments hoping for finite subcovers of general open covers, and left disappointed....
posted by pjm at 10:44 PM on February 11 [2 favorites]


The definitive guide to the REDMAP story is the book Ratf**ked (2016), which I finished reading this afternoon. Subtitled “The True Story Behind the Secret Plan to Steal America’s Democracy.“

Among the classic contorted ’ink blot‘ districts is Pennsylvania’s 7th, described as looking like “Donald Duck kicking Goofy.”

The unsung ‘hero,’ if that’s the right word, of the REDMAP projecct was a little-known man behind the scenes named Chris Jankowski.
posted by LeLiLo at 12:43 AM on February 12 [1 favorite]


The map generated by compactness algorithm feels all wrong, [...is...] basically just a case of Engineer's Disease

I don't think it counts as Engineer's Disease: it was only intended to illustrate how bad the problem is, not to say "here, I have solved your problem with some coding".
posted by ambrosen at 2:13 AM on February 12 [1 favorite]


Why not keep single-member electorates, but dump their connection to geography? Keep everyone in the electorate they're in at present, but assign immigrants and other new voters to a randomly-chosen electorate. Over time the effects of gerrymandering will be eroded and you'll end up with fairly balanced electorates.
posted by Joe in Australia at 4:11 AM on February 12


Why not keep single-member electorates, but dump their connection to geography?

The majority of people don't move around a lot, so you'd still have that connection to geography, except the people who did move away would have essentially no representation as well.
posted by Etrigan at 6:00 AM on February 12 [1 favorite]


As seen on MR, most of the partisan problem may be due to actual voter clustering. I'm not sold on reverse gerimanders which try to make pr happen at the cost of meaningless districts, nullifying one of the few reasons for the system. in a real way fptp systems will always leave substantial fractions of the population of no interest to their representatives.
posted by a robot made out of meat at 6:16 AM on February 12 [3 favorites]


Here's another reason to be concerned about gerrymandering:

Republicans in Minnesota, Virginia propose changes to their electoral college rules

Right now only Maine and Nebraska do this. If a couple of big swing states do this, you will rarely see Democratic presidents again.
posted by great_radio at 9:05 AM on February 12 [4 favorites]


I've said this in pretty much every gerrymandering thread, but to avoid the ills of gerrymandering you don't want pretty compact districts or any other such thing as is usually brought in as the golden criterion in these discussions.

The reason (as pointed out already above) is that voters of any given political party tend to cluster geographically. So it's very possible and in fact easy to make super-nice looking compact contiguous districts that are in fact gerrymandered as anything.

FWIW here in Missouri our state senate districts are required to follow county lines whenever feasible. That system turns a 55/45 Republican majority into about an 80/20 majority in the state Senate.

This type of thing is bad for democracy and the fact the most R & D voters sort themselves by county doesn't make it any better.

In determining what is or is not gerrymandered, you need to look at two vitally important criteria:
  • Does the distribution of representatives elected from given districts reasonably approximate the distribution of the voting population as a whole? For example, if voters are 60/40 R/D then we shouldn't see elected reps 80/20 just because of the way the districts are drawn.
  • Do we have as many competitive districts as possible? "Safe" districts are bad for democracy. It doesn't matter if they are safe D or safe R.
posted by flug at 1:42 PM on February 12 [6 favorites]


I will mention that some (many?) democracies solve this type of problem by assigning some reps to parties based on the proportion of the total vote that party has received. This helps solve both issues I mentioned above (distribution of reps far different distribution of voters & non-competitive districts).

I will also mention that Ds often see solving the gerrymandering problem as key to their future success. Spending some time & effort on this is worthwhile but I will suggest something far more profitable: Figuring out a platform, strategies, candidates, etc etc that will allow them to win despite the current districting issues that seem to be somewhat tilted against them.

Some of that tilting has indeed been purposefully put in place by their opposing parties and some degree of it may be un-doable. But a ton of it (ie, the geographical sorting by state that naturally tilts the U.S. Senate & Electoral College) is anchored deep in the Constitution, was put in place that way on purpose many years ago in order to slightly favor certain states and regions of the country, and is not going to be easily changed.

Political parties need to learn how to win given the currently political landscape, and that includes the specifics of districting. If you can't win given the hand your are dealt to play with, you can't win period.

Democrats can and should be winning for more seats even given current districting. Why are they not? That is a question that party leaders and members should be spending a lot more time thinking about.
posted by flug at 1:56 PM on February 12 [1 favorite]


It's not a good idea to eliminate legislative districts - because geography matters.

I think we could (hypothetically) have one of the two legislative branches awarded proportionally ... but that would require changing the way we do government in many ways. It would mean we wouldn't be electing people, just parties, and I'm pretty sure many people wouldn't be happy about that.

Democrats can and should be winning for more seats even given current districting. Why are they not? That is a question that party leaders and members should be spending a lot more time thinking about.

They're not, in part because the leftists who think Democrats should be winning wouldn't be happy with the representatives we'd get. Leftists keep saying that Democrats should move farther left -- but, if we want to win in conservative districts, we're going to go much farther right.

Which I'm actually okay with. Idealism (on the part of a handful of voters) cost us the 2016 elections. But those self-same leftists who bash Democrats for not winning aren't going to be happy with the representatives we get if we use a 50-state strategy in 2018.
posted by steady-state strawberry at 2:12 PM on February 12 [1 favorite]


I think we could (hypothetically) have one of the two legislative branches awarded proportionally ... but that would require changing the way we do government in many ways. It would mean we wouldn't be electing people, just parties, and I'm pretty sure many people wouldn't be happy about that.

That's how things work here, in Australia, and the results haven't been great. You tend to end up with two major parties/coalitions dominating the chamber, and the deciding votes going to a few "independent" parties who have disproportional influence.

It's even worse in unicameral systems like Italy and Israel's, where every government is almost necessarily a coalition that relies on the most extreme minor parties to make its majority. What the major parties should do is form a coalition themselves, but that's obviously too sensible for anyone to agree to.
posted by Joe in Australia at 2:17 PM on February 12


Since we limited the number of representatives at 435, the districts are now so large that it is practically impossible to create a district that has any sort of cohesion on geographic

In theory smaller districts should enable a for more representative selection of elected representatives.

But I will just give you the case of Missouri as a counterexample. We have 8 Congressional Districts, 34 State Senate Districts, and 163 State House Districts. 34 is more than 8, and 163 is a LOT more--so each should be proportionally more representative than the last, right?

Here are the actual results for Missouri:

* Distribution of the voting population: 56.9% Republican - 39.7% Democratic *
* US House seats: 75% Republican - 25% Democratic
* State Senate seats: 72.7% Republican - 27.3% Democratic
* State House seats: 72.0% Republican - 28.0% Democratic

First off, you will note that each of the three districting solutions gives Republicans about 27%-33% advantage (depending on exactly how you count it**) over their popular vote totals.

Second, you'll notice that there are slight differences among the three elected bodies--but these really amount to just rounding errors (with only 8 U.S. House districts, each seat represents 12.5% of the total, so you can't cut things any finer than that. State Senate has 34 seats so each Senator represents about 3% of the total, can't cut any finer than that, etc). Given those constraints, the proportion of the three chambers is pretty much as identical as it could be made. So there is clearly something very consistent happening here, even though the three sets of districts are quite different.

Conclusion is that simply adding more districts doesn't really do anything fundamental to solve the problem. It's not how many districts, but how they are drawn that counts.

* FYI this is an average of R/D votes for five statewide races in the Nov 2016 election.

**Republicans won the popular vote by 17.2%. They won the three districted chambers by 43.9%, 45.5%, and 50%, respectively. So simply lumping voters into districts shifts the margin by 26.7%, 28.3%, and 32.8%, respectively. That is . . a lot.

posted by flug at 2:26 PM on February 12 [2 favorites]


That's how things work here, in Australia, and the results haven't been great. You tend to end up with two major parties/coalitions dominating the chamber, and the deciding votes going to a few "independent" parties who have disproportional influence.

It's even worse in unicameral systems like Italy and Israel's, where every government is almost necessarily a coalition that relies on the most extreme minor parties to make its majority. What the major parties should do is form a coalition themselves, but that's obviously too sensible for anyone to agree to.


See, I'm not sure that either of those possibilities is worse than what we have in the U.S. right now. You have one party, the minority party by many standards of counting, which somehow is dominating everything and doesn't have to make a coalition or compromising with anyone or anything* in order to simply rule.

If they even had to compromise with some very small party comprised entirely of nutjobs, that would be an improvement over the current situation, where they literally don't have to compromise with anyone at all.

* Well, except for themselves, of course. We'll see how that goes over the next, say, 3.9 years . . .
posted by flug at 2:35 PM on February 12 [2 favorites]


Does the distribution of representatives elected from given districts reasonably approximate the distribution of the voting population as a whole? ... Do we have as many competitive districts as possible?

These two things are at absolute loggerheads. Maximizing competitive districts is exactly the same thing as gerrymandering in favor of the local minority party.
posted by ROU_Xenophobe at 2:36 PM on February 12 [3 favorites]


Part of what makes this such a complicated problem is you can't design the system to assume party alignment is invariable, or that just because a particular population center is mostly registered Dems or registered Republicans, that means we can assume they wanted their party's candidate to win. It may seem incomprehensible to some of us, but Dems sometimes vote for Republicans and vice versa, because party preference isn't actually a demographic feature, it's not legitimately a part of someone's identity. It's a voluntary group affiliation or affinity at a point in time, not a fixed property of individual people. Who doesn't know tons of other people whose party affiliations have changed or who have crossed party lines to vote for the other side? I do. Maybe that's less common than it used to be now due to the hyper partisanship we've been seeing, but is it really a good idea to base anything about how you draw districts on assumptions about party affiliation being relatively fixed or stable?
posted by saulgoodman at 4:03 PM on February 12 [1 favorite]


Saulgoodman, the bad guys do make that assumption, and it has worked for them. Gerrmandering may not always work, and it's theoretically susceptible to landslide losses, but it has been pretty successful at ensuring Republican victories across the USA.
posted by Joe in Australia at 5:51 PM on February 12 [2 favorites]


>> Does the distribution of representatives elected from given districts reasonably approximate the distribution of the voting population as a whole? ... Do we have as many competitive districts as possible?

> These two things are at absolute loggerheads. Maximizing competitive districts is exactly the same thing as gerrymandering in favor of the local minority party.


Ah yeah, no.

There is indeed some degree of tension between those two goals. In a typical situation you cannot (for example) make every district extremely competitive.

But let's look it what happens in a typical gerrymander situation: You end up with a few 80-20 or 90-10 districts for the minority party. And you end up with a LOT of 55-45 or 60-40 districts for the majority party.

That is how you end up with:

#1. Vastly more districts favoring the majority party than their proportion of voters would indicate

#2. NO competitive districts for either party

So to counteract the caustic effects of gerrymandering, you have to negate BOTH of those effects.

You're right that you probably won't end up with competitive races in 100% of districts in most situations. But you can certainly end up with a situation where 50-60-70% of races are within 5% or so--close enough that the candidate feels the heat every election.

Instead of 0% of races competitive, which is the situation in most heavily gerrymandered areas.

And yes, it's possible to have many more competitive districts than we currently do while still approximating the representation of political parties in the general population in the outcome.
posted by flug at 7:43 PM on February 12 [2 favorites]


Maximizing competitive districts is exactly the same thing as gerrymandering in favor of the local minority party.

And I'm going to say it again--more emphatically if I can figure out how:

The term "gerrymandering" focuses our attention on the shape of the districts, and somehow we arrive at the conclusion that strange-shaped districts are bad.

The problem with gerrymandering is not the shape of the districts.

The problem is the corrosive effects that gerrymandered districts have on our political system. The corrosive effects are precisely that
  • One party gets a (often massively) larger representation than they deserve based on their popular support, and
  • Competitive districts are minimized or even eliminated
If we want to stop the ills of gerrymandering, we must stop focusing on the shape of the district, and focus directly on the political ills gerrymandering causes.

Your opponents are thinking about gerrymandering all the time. And I guarantee you, they are not thinking at all about pretty shapes. They are thinking about what political advantage they can gain by gaming the process.

They are analyzing, carefully and in detail, voting patterns by geography. You are never going to defeat them or take away their advantage by ignoring that factor and concentrating on things like pretty shapes.

And if you set up a criterion, like "must have the most compact shape possible" or something--then they will just use the powerful analysis tools at their disposal to generate a "pretty shape map" that accomplishes their political goals just the same as before.

If you want to stop parties from using districting to accomplish this kind of political goal, you have to frankly admit that they are doing this and directly address the goals they wish to achieve by putting different goals in place.
posted by flug at 7:52 PM on February 12 [4 favorites]


That's true, but if you actually bake those assumptions into the system, we make mind reading part of the process. No one is or should in any way feel that their allegiance to a party commits them to supporting that party or assumes that alignment. Fundamentally, voters aren't supposed to be loyal to a party in the constitutional design of our system. We have parties now due to a quirk of history--a dispute that fomented into a split we've had since almost the beginning not within the constitutional design itself, but external to it. The party system has some legal basis now, as there are rules and laws that circumbscribe party political activities and fund raising, etc., but none of that body of law has any specific constitutional basis in the U.S.

Washington and the other founders weren't total idiots. They knew about political parties because those already had historical precedent in Europe, but many (like Washington) specifically believed political parties were bad for systemic reasons--namely, that the parties themselves would be in perpetual competition when they were explicitly aspiring to design a system that would be less fractious and more cooperative, and because political partisanship had so often provided cover for outside networks of political influence to interfere in European nations' internal political affairs to the detriment of the people. That sort of outside interference and loyalty to party before country--the kind that had already been seen in the British and other European systems--was something they wanted to avoid allowing to develop in our system.

But let's face it: Jefferson was a hothead who spoke glowingly and reverently of the need for periodic blood baths as a kind of purification ritual or sacrifice to the dark gods of liberty. He was a puzzle wrapped in an enigma, as the saying goes. Washington was a much more self restrained, judicious kind of leader, a lot like Eisenhower, as statesmen go. He may have been on the right side of history and we're really just now starting to see the longer term schism caused by that early split that first established our party system. There were no checks and balances designed to check those parties, regardless of whether you believe they serve some vital purpose or are indispensable or whatever. It's not the case that any part of our constitutional design was ever conceived or expected to function as a partisan system. It absolutely was not, whatever else might be true of parties in our system now.
posted by saulgoodman at 7:59 PM on February 12


These two things can't be true at the same time:
There were no checks and balances designed to check those parties
versus
It's not the case that any part of our constitutional design was ever conceived or expected to function as a partisan system.
If there was nothing in the design to prevent political factions from cohering into parties and wielding power and pursuing goals in a coordinated fashion, then it doesn't matter what they were conceiving or expecting: it is fundamental to the design of the system that there are going to be political parties. The designers were just shitty at that aspect of designing, conceiving, and expecting if indeed they didn't expect it to happen at all.

(Which is entirely understandable: it's not like there was a textbook or a Journal of Founding Perpetual Ideal Democratic States where they could simply read up on it, at least nothing practical and evidence-based, and book learning about it would hardly be adequate anyways but there also wasn't any sort of apprenticeship program where you could polish your skills and get the kinks worked out while founding your first dozen nations, either.)

Simply declaring "There should be no parties!" makes no more sense than saying "There should be no career politicians! Everyone should just be like Cincinnatus and serve when called and cede power unhesitatingly when the need for wielding that power is resolved!" when you don't actually put term limits in place or other measures to impair the practicality or utility of a career in politics, or otherwise systemically dissuade people from trying to hang on to power.

(Note that I'm not saying that we should have term limits or should not have career politicians, just that there definitely are going to be career politicians if you do absolutely nothing to stop it from happening, and it's going to be false to assert that the system of its nature abhors career politicians, regardless of ideals or expectations.)

Also note that not only are political parties inherent in this design, but scholarship during the 20th century showed that specifically, two major parties is the steady state/attractor of a FPTP system.
posted by XMLicious at 9:34 PM on February 12 [3 favorites]


Instead of 0% of races competitive, which is the situation in most heavily gerrymandered areas.

If 0\% of the races are competitive, then by definition it is not gerrymandered in a partisan fashion. Or, I suppose, people were trying to and really fucked up. If you are creating uncompetitive districts for your own party, you are wasting votes. You are ceding districts to the other party -- the people who have terrible ideas, the people whose policies would hurt the country.

If you are maximizing competitive districts, you are maximizing the number of districts in which the local minority party might win. This is the same thing as gerrymandering in their favor. Or, really, not even the same thing, it just is gerrymandering in their favor.

The term "gerrymandering" focuses our attention on the shape of the districts, and somehow we arrive at the conclusion that strange-shaped districts are bad.

Okay, but quoting my earlier comment to say that seems weird, because I didn't and wouldn't mention weird shapes.
posted by ROU_Xenophobe at 10:01 PM on February 12


scholarship during the 20th century showed that specifically, two major parties is the steady state/attractor of a FPTP system.

Sort of, except that it mostly doesn't work (ie, the UK, Canada). It's more Duverger's Sort Of Vaguely Applicable General Principle With Loads Of Caveats And Exceptions And Yeah-Buts.
posted by ROU_Xenophobe at 10:05 PM on February 12 [1 favorite]


Since neither of us referred to it as a "law", but that seems to be what your phrasing is opposed to, you may be arguing with yourself.
posted by XMLicious at 11:29 PM on February 12


True dat.
posted by ROU_Xenophobe at 6:04 AM on February 13


it is fundamental to the design of the system that there are going to be political parties.

That's not how the people who designed our system saw it. The fact they intended not to have parties and saw them as dangerous isn't really debatable. That much is in their own words. That's as far as I'm claiming, except to say, I think they were probably right.
posted by saulgoodman at 8:43 AM on February 13


"For where two or three people gather in one name, there a party is with them."
posted by tobascodagama at 9:19 AM on February 13 [1 favorite]


They meant the kind of party that has a bank account and a fictional name license, a legal structure for pooling money when they spoke about parties at the beginning. Not organised movements, but businesses designed to get specific candidates elected.
posted by saulgoodman at 9:22 AM on February 13


I mean, maybe it was dumb and impractical, but the idea was to let ideas not people be the main show in our system. So I don't think organizations promoting different ideas would have been out of the question if that original plan had worked out. Only formally incorporated entities working to pick and elect specific people instead of lobbying the public to support different ideas and the candidates willing to pledge to adopt them.
posted by saulgoodman at 9:27 AM on February 13


Are we still talking about something like what they "knew about political parties because those already had historical precedent in Europe"? Were these precedents "the kind of party that has a bank account and a fictional name license, a legal structure for pooling money... Not organised movements, but businesses designed to get specific candidates elected"?

The page you linked to there, by the way, says
The leaders of the American Revolution did not like the idea of parties and political battles between parties. Upon his retirement from public life in 1796, George Washington warned Americans against "faction" (parties). James Madison thought parties were probably necessary, although he did not entirely approve of them. Alexander Hamilton thought that faction was a vice to be guarded against at all times. Thomas Jefferson declared in 1789, "If I could not go to heaven but with a party, I would not go there at all." Nevertheless, the men who held these views founded the first two great American political parties.
It would be especially dumb and impractical to design a system to not have political parties and then go found political parties yourself. But I think you may be attributing more definite beliefs and guidance, or at least more synoptic and non-contradictory beliefs and guidance, to the people who laid down the political system than they actually held or gave.
posted by XMLicious at 9:57 AM on February 13


Their letters say otherwise. They did discuss this. The more idealistic plans just went awry because they got into a power struggle.
posted by saulgoodman at 10:17 AM on February 13


And you know, chartered organizations existed before the revolution. The founders mostly came from gentry. They had the practical experience to think about it in terms of systemic, organizational issues. They also rejected a formally organized, standing military for similar reasons. They had a lot of protoanarchist revolutionary ideas about organizations and systems at first. Those ideals didn't necessarily make it far in practice, but as a thought exercise, I'm still having a hard time seeing what other than corrupting influences and more limited ranges of policy options the party system gives us.
posted by saulgoodman at 10:23 AM on February 13


In another thread, when I asked you how you would get rid of parties, at first you said doing nothing would be adequate... but now I'm noticing that in a subsequent answer you seem to be saying that candidates who didn't satisfy non-partisan criteria just wouldn't be allowed to run for office, which is about 180° opposite from "nothing".

So I guess I have to apologize: in my comments above and in the other thread I thought that you were still maintaining that something designed into the system by the drafters of the Constitution would prevent new parties from forming if we abolished the existing ones. Sorry about that, the first time I read through I missed the most relevant detail.
posted by XMLicious at 11:02 AM on February 13


Or, reading it a couple more times I'm still not so confident I understand what you're saying.

In any case, though, my point isn't that parties are somehow virtuous, it's that you would have to have a system radically different from the one specified in the Constitution to not have any parties.
posted by XMLicious at 11:10 AM on February 13


That's not how the people who designed our system saw it. The fact they intended not to have parties and saw them as dangerous isn't really debatable. That much is in their own words.

Sure, to some extent. But you should not care.

You should not care what they thought about parties any more than you would care what they thought about black people, or women, or homosexual people. You should not care what they thought about parties because the vision they had for life in the new United States is one that you would be rightfully repelled by. Government the way the framers intended is something no decent person could possibly support.
posted by ROU_Xenophobe at 11:39 AM on February 13 [3 favorites]


Honestly, as a system's engineer, my intuition on this is the same as Washington's. I really don't see any value in the parties at all.
posted by saulgoodman at 5:11 PM on February 13


Aren't there a bunch of right-wing state legislatures pushing for a constitutional convention right now? This is a Republican effort. Texas. Tennessee. Utah. But not South Dakota.

It's kind of amazing how the GOP has locked up so many avenues of political power. The White House, Congress, the Supreme Court, and state-level legislatures and governorships. Why is this happening now? Amending the Constitution by Republican-controlled state governments under the rule of a Republican-dominated federal government is just piling insult upon insult.

Has anyone on MeFi been talking about this proposed conservative-driven constitutional convention of states? Oh wait I guess there was some comments back in December. But I recall asking why not repeal the electoral college in early November and being blithely told that it was impossible. But the Republicans are trying to do it now...?
posted by Apocryphon at 11:41 PM on February 13


Has anyone on MeFi been talking about this proposed conservative-driven constitutional convention of states? Oh wait I guess there was some comments back in December. But I recall asking why not repeal the electoral college in early November and being blithely told that it was impossible. But the Republicans are trying to do it now...?

My impression is that the constitutional convention push is a separate effort from electoral college reform. The convention is to propose stuff like "religious freedom" amendments, defining marriage as between one man and one woman (probably with transphobic provisions for good measure), federal term limits, etc.

Republicans don't want to get rid of the electoral college, they want to ratfuck it at the state level by having large swing states allocate electoral votes by (gerrymandered) districts. For example, Obama won Pennsylvania by five points in 2012, but only carried five congressional districts. Under the Republicans' proposed allocation, Obama would have received only seven of the state's 20 electoral votes.
posted by HumuloneRanger at 8:20 AM on February 14 [1 favorite]


Well exactly, the currently proposed constitutional convention is to push right-wing agenda, probably a balanced budget amendment in addition to the other policies you mentioned. My point is, one side is still dreaming wistfully at the impossible day the electoral college can get amended out, and the other side is already on the way to getting a convention of states to rewrite the Constitution as they see fit. Shouldn't people who don't want this country to be rolled back to centuries past be more concerned about this? Why is there no discussion about preparing to deal with this?
posted by Apocryphon at 9:13 AM on February 14


My point is, one side is still dreaming wistfully at the impossible day the electoral college can get amended out, and the other side is already on the way to getting a convention of states to rewrite the Constitution as they see fit. Shouldn't people who don't want this country to be rolled back to centuries past be more concerned about this? Why is there no discussion about preparing to deal with this?

Amending the Constitution requires a super-majority in either congress or at the state level; winning elections is the way to deal with that, so when you hear people talking about 2018 or 2020 it's not really a separable discussion -- this is something I'm idly worried about, but it requires the same solution as Republican legislation does generally. And because only a fraction of seats and offices are up for election in any given year, it is (I think, offhand, but this is easily checked) actually numerically impossible for the Democratic party to obtain the necessary majorities to pass and ratify amendments without Republicans assistance in 2018-19, even if we assume a Democratic sweep in 2018; but conversely, it's very possible to block Republican chances of passing an amendment or calling for a convention. Strategically, I think there's not a lot of energy around calling for an amendment to end the electoral college when that's going to need to be a bipartisan movement if it's to happen before 2020 -- if 2018 goes well, that's a conversation that might start happening, particularly as the 2020 Presidential election draws closer.
posted by cjelli at 9:57 AM on February 14 [2 favorites]


Why is there no discussion about preparing to deal with this?

Not much needs to be done. Even if there were a convention that came up with dumbshit stuff, all it can do is propose amendments to the states. Nothing would actually change unless 38 states ratified the amendments, which would be difficult for creepy right-wing stuff. Playing with Gallup numbers, you only need states at least as liberal as Illinois to reject them and they're dead.
posted by ROU_Xenophobe at 10:28 AM on February 14


Not much needs to be done. Even if there were a convention that came up with dumbshit stuff, all it can do is propose amendments to the states. Nothing would actually change unless 38 states ratified the amendments, which would be difficult for creepy right-wing stuff. Playing with Gallup numbers, you only need states at least as liberal as Illinois to reject them and they're dead.

Yup. Democrats control both houses in twelve state legislatures (HI, CA, OR, NV, NM, IL, MD, DE, NJ, RI, MA, VT) and there are another six states that are split (AK, WA, CO, NY, CT, ME). As cjelli says, the important thing to do is to get Dems - or other progressives - elected in as many positions as possible.

As for eliminating the electoral college, you need approval of at least 38 states. Seven states (plus DC) have outsize representation in the EC with three EV each; another five states have four EV each. Then there are the swing states (FL, PA, OH, NC, VA) that will be loath to give up their influence.
posted by HumuloneRanger at 3:34 PM on February 14


"Has anyone on MeFi been talking about this proposed conservative-driven constitutional convention of states? "

A whole bunch of state constitutions require voters to vote every 10 or 20 years on whether they want a state constitutional convention (popularly, "con-con"). There's always someone pushing for a con-con, every election -- sometimes because they're true believers, sometimes because they're trying to drive turnout on unrelated issues. But the professional politicians in the statehouse basically never actually want a con-con (aside from a few true believers) because the things in the state constitution that the state legislature wants changed bear very little resemblance to the things the people in the state want changed.

Illinois voted on its con-con in 2008 and there was actually a fair movement in its favor because of our pension problems (which the state constitution addresses and prevents many solutions to), but still nobody wanted to open the hornets' nest of having an actual con-con where everything's up for grabs rather than just seeking specific, individual amendments.
posted by Eyebrows McGee at 5:55 PM on February 14 [3 favorites]


As for eliminating the electoral college, you need approval of at least 38 states.

For the amendment, yes, but you could in theory do it with less via the national popular vote compact. Similar problems though - most of the votes are in the important swing states, and they're going to be loathe to sign on. Right now we've got CA and NY, which is a good chunk right there, but it's hard to see what other state with more than 10 EC votes is going to sign on.

It needs another 95 to go into effect. If we take all the Democratic controlled legislatures that gives us 21, and if we somehow got the split states that's up to 44, and we'd still need states worth another 51 EC votes for it to happen.

That said, unlike (modern) proposed amendments, the national popular vote compact doesn't expire, and who's a swing state does change over time. I can imagine a scenario where the new swing state doesn't manage to repeal the compact once signed, and a former swing state decides to sign up now that they're less in play. But that's decades into the future, and little help to us now.
posted by vibratory manner of working at 6:33 PM on February 14


That said, unlike (modern) proposed amendments, the national popular vote compact doesn't expire, and who's a swing state does change over time.

Unlike an actual amendment to remove the electoral college, the national popular vote interstate compact doesn't actually remove the electoral college: it binds the electors of those states to vote as the popular vote dictates. That might seem like a minor distinction, but it does mean that there issues with the electoral college -- for example, faithless electors -- that remain under the compact.
posted by cjelli at 8:42 PM on February 14


A whole bunch of state constitutions require voters to vote every 10 or 20 years on whether they want a state constitutional convention (popularly, "con-con").

What the hell, that sounds like a dead man's switch that dissolves the union and reverts back to the Articles of Confederation
posted by Apocryphon at 11:24 AM on February 15


Eyebrows McGee is talking about conventions discussing each state's constitution, not the U.S. Constitution.
posted by XMLicious at 11:47 AM on February 15


« Older Measurement error and the replication crisis   |   Hopper College! Newer »


This thread has been archived and is closed to new comments