One weird trick to end gerrymandering: cake-cutting game theory


#1

Originally published at: https://boingboing.net/2017/11/06/one-weird-trick-to-end-gerryma.html


#2

IMHO CGP Grey’s videos on this topic are a must see.

I think these three are the most relevant ones. Sorta surprised he didn’t make a playlist.


#3

It won’t work because the representatives of the safe districts for the losing party (the two all blue districts in the example) are very often completely complicit in the gerrymandering; sacrificing the good of their party and constituents to avoid having to work to defend that seat.


#4

What incentive does the party in power have for adopting this procedure as opposed to the current practice?

I suppose if the alternative was something completely outside their control, like applying some statistical clustering technique on the portion of the census data that can legally be taken into account when determining voting districts, that might do it.


#5

This assumes that anybody with even a shred of power wants fairness.


#6

From the first video:

Independent committees are still not ideal because they usually group similar areas together, so the elections are usually uncompetitive.

Sometimes I get frustrated thinking about gerrymandering, because it’s not always clear what should be the correct result. Isn’t the point of a representative that they represent a group of people with similar needs? So what’s wrong with grouping similar areas together?

I get that an uncompetitive district can lead to stagnation, but one solution to that is term limits.

The other thing I don’t get is that it requires a division of congress by geographic bounds, and yet we’re hoping for a balanced two-party congress that matches political views of the constituents. But this can only work if the voters always cram themselves into districts with like-minded people.

If, instead, you had a state where 60% liked Party A, and 40% liked Party B, but the voters were all distributed evenly in the state, then Party A would always win 100% of the seats, no gerrymandering required.

I’m not sure what the solution is, but I wonder if you couldn’t have a system that does away with the strict geographic boundaries, and instead allows groups of 700,000 people across the state to link with each other and elect their own representative. Then you could have representatives that truly represented a group of like-minded people, not just people in a neighborhood.


#7

MMP


#8

Why do we still insist on dividing up representation by location of residence? This made sense back in an era when most people were tied to large farms but things have changed. A lot of people spend most of their time (and money) in other districts. Wouldn’t allocating legislative representatives to like-minded people be better?


#9

Moon Duchin, Associate Professor of Mathematics, Tufts University, who is speaking at this conference is also the convener of a group of mathematicians and geometers who are working on the issue of gerrymandering.

Conference on Redistricting Reform at Harvard University
Wednesday, November 8
8:30 AM – 7:30 PM EST
Harvard, Ames Courtroom, Austin Hall, 1515 Massachusetts Avenue, Cambridge
RSVP at https://www.eventbrite.com/e/conference-on-redistricting-reform-at-harvard-university-tickets-37853844869

You are cordially invited to join the Ash Center for Democratic Governance and Innovation for a for a day-long conference on gerrymandering, redistricting, and the fight for American democracy. As the Supreme Court prepares to hear and hand down its decision in the most important case on political gerrymandering in a generation, join us for a convening of scholars, practitioners, and activists from around the country, as we seek to understand a path forward on redistricting reform. Our panelists will provide unique insight into the history of gerrymandering, the role of race in redistricting, the current crisis of the Census, and examine the landscape for reform.

The day long program will close with a public talk at the American Academy of Arts and Sciences from 5 - 7:30 pm. A draft agenda for the day is below.
Agenda
Date: November 8, 2017
Location: Ames Courtroom, Austin Hall, Harvard Law School
8:30: Coffee and light breakfast
9:15: Welcome
Randall Kennedy, Michael R. Klein Professor, Harvard Law School
Douglas Elmendorf, Dean, Harvard Kennedy School
Archon Fung, Academic Dean and Ford Foundation Professor of Democracy and Citizenship, Harvard Kennedy School
Miles Rapoport, Senior Practice Fellow, Ash Center for Democratic Governance and Innovation
9:45: Gerrymandering: A Tortuous History
Alex Keyssar, Matthew W. Stirling Jr. Professor of History and Social Policy, Harvard Kennedy School
10:45: Race and Redistricting
Randall Kennedy, Michael R. Klein Professor, Harvard Law School, Moderator
Terry Ao Minnis, Director of Census and Voting Programs, Asian Americans Advancing Justice
Janai Nelson, Associate Director/Counsel, NAACP Legal Defense Fund
Thomas Saenz, President and General Counsel, Mexican American Legal Defense and Educational Fund
12:00: Redistricting Through a Partisan Lens (lunch to be provided)
Chris Jankowski, Republican Consultant and Former Director, Operation REDMAP
Kelly Ward, Executive Director, National Democratic Redistricting Committee
1:30: The Landscape of Reform: Issues and States
Moon Duchin, Associate Professor and Director, Program in Science, Technology, & Society, Tufts University
Cathy Duvall, Redistricting Reform Project
Dan Vicuña, National Redistricting Manager, Common Cause
Michael Li, Brennan Center, Senior Counsel, Brennan Center for Justice
Wendy Underhill, National Conference of State Legislatures
3:30: The Crisis of the Census
John H. Thompson, Executive Director, Council of Professional Associations on Federal Statistics and Former Director of U.S. Census Bureau
Arturo Vargas, President, National Association of Latino Elected Officials
Location: House of the American Academy of Arts and Sciences (shuttles provided).
5:00: Reception
6:00: AAAS “Stated Meeting”: Redistricting & Representation
Moon Duchin, Associate Professor of Mathematics, Tufts University
Jamal Greene, Dwight Professor of Law, Columbia Law School
Gary King, Weatherhead University Professor and Director of the Institute for Quantitative Social Science, Harvard University
Moderator: Hon. Patti Saris, Chief Judge, U. S. District Court, District of Massachusetts


#10

So why exactly can’t each of the states get a number of seats and distribute them the cascading or weakest link way?

Cascading election: each voter names X candidates, where X is the number of votes. X candidates with most votes win

Weakest link: Choose X candidates in order from the one you think most deserves the job. Assign each candidate as many votes as they got, then eliminate the weakest link and move his voters to their second choice. Repeat until X candidates are present, where X is the number of seats to fill


#11

Here’s the Vox take


#12

The problem with this is that then you’d have no representation at all for the less densely-populated area of the states. For example, New York state would probably see all of it’s reps come from the 4 largest boroughs, while Buffalo (despite being a pretty large city in the US) would have approximately zero representation.


#13

Great! The entire state’s representation is from a single party. It combines all the worst features of first past the post with all the worst features of a party list system. Might as well just have a single representative, and vary the number of votes they have.


#14

I think you could make a ranked or jungle vote system work, but you’d have to do more work to tabulate the results than simply take the top X vote-getters.

First, everyone votes for the candidate they want. As many candidates can run for each party as they like. Once the vote is in, the party vote total is used to determine how many representatives from that party can be seated in Congress. If a state with 10 districts votes 60% Democratic, 30% Republican, and 10% assorted third parties, then the top 6 Democratic candidates are seated, followed by 3 Republicans, and finally the top-voted member of the remaining parties. Even if 30 Democrats or Republicans run, they can’t crowd out the rest of the parties by sheer numbers.


#15

Don’t have enough likes for this.

When politicians are making districts for political reasons, you don’t live in a democracy. Democracy is the people choosing their leaders, not the leaders choosing their people.

It’s so easy if you use the greatest voting system of all: Random Ballot!

Just have everyone cast their votes, then pick a ballot at random. That person wins. If you need to pick five winners, just keep picking ballots.

It’s seriously the best.


#16

So a variant of Asimov’s “Franchise”?


#17

Not that it’s much different than simply picking the N candidates with the highest totals, just the addition of luck.

Right now I’m reading up on proxy voting to see what weaknesses it has.


#18

The Supreme Court might make them.

http://www.npr.org/2017/10/03/555425809/supreme-court-set-to-consider-partisan-gerrymandering


#19

I’ve always thought a significant crimp could be put into this nonsense by simply creating ratio of perimeter to area that cannot be exceeded. It would put a stopper on these 40-mile 30 foot wide necks along a roadway to another demographic clump and other tortured geometry.


#20

In California we have since 2010 a Citizens Redistricting Commission that is in charge of drawing districts based on a fixed set of rules. It uses a process that includes public input and open hearings. Not perfect, but better than having the legislators draw their own districts.