Proportional, Accountable, Local (PAL) representation: isn't this a big deal?

82 views
Skip to first unread message

Jameson Quinn

unread,
Oct 25, 2011, 8:32:40 AM10/25/11
to EM, electionsciencefoundation
Yesterday I posted two messages to the Election Methods list about PAL representation, a new PR system designed as a replacement for single-member districts. The system combines elements from STV, Balinski's "Fair Representation", and SODA. 

Basically nothing in the system as I described it was in itself a new idea. My goal was not to create an innovative system, but to blend existing ideas to find a proportional system that would be acceptable to all existing interests, including voters, theorists/activists like us, incumbent politicians, and parties.

I believe that a system like this is the only kind of PR that could pass in a US environment, and also well-suited to the UK. I'm a little discouraged that a development I regard as truly important got no comment from the list.

So I'm writing another message. My excuse is that I've thought of one additional rule for the system, a way to make a party threshold that's fair to independent candidates. Also, I've expanded and rewritten the article (included at bottom here).

Why could no other system PR system pass in the US? (These arguments also might apply to other single-member-district countries such as the UK or Australia, but I can't say for sure.)
  • A closed list system would be (rightly) attacked as a power grab by party bureaucrats. Voters have been souring on parties for decades now, and they wouldn't stand for that.
  • A global open-list system such as STV would have unacceptably-complex ballots. Who can keep track of dozens of candidates, let alone fully rank them?
  • A districtless system would be too radical a change. People are used to having "their" representative.
    • Note that PAL has one representative per party for each district. Each representative is associated with multiple districts. So your representative is the one from your party from your district. If you voted for a party that got no seats, your vote was probably transferred to a "backup" party according to a preference order you assented to, so you still probably have a representative. ("probably" because up to a Droop quota of voters have no representative.)
  • A multimember-district system helps with the above problems, but doesn't actually solve them. Who wants a system where ballots are only a little bit too complex, where you only sort of know who your representative is, and which is only mostly proportional?
  • A mixed member system is an ugly hybrid. To me, US democratic ideals are too egalitarian to accept that there could be two different kinds of representative.
  • More seriously, a mixed member system would be totally unacceptable to existing incumbents, as it would draw too many of them out of their existing districts. And I actually think this is in part a valid concern. The public interest is to have representatives who are accountable, not complacent; but I see no value in change for change's sake.
  • Balinski's "Fair Majority Voting", as used in Belgian municipal elections, resolves all of the concerns above, but it would be very hard to justify the fact that some representatives would lose with a majority vote. It's very hard to respond to a simple question like "Why should my opponent win with 45%, when I lose with 52%?" with a complex answer about party balance and compensating for gerrymandering.
    • Note that PAL representation would actually give the same result as FMV, but would provide an easy justification for that result. Responding to the question above, you could say: "Each representative needs exactly the same number of votes to win. Your opponent got the vote transfers they needed to reach that threshold and you didn't. Those votes were transferred in accordance with the explicit will of the voters, and to ignore them would be to disenfranchise those voters."
(Random note: I did the math for how many people in the US would be part of the unrepresented leftover droop quotas in a state-by-state PR system. The answer is around 20%. If you assume that most elected representatives got a bit more than the quota, so that the unrepresented fragment is only 2/3 the size of the represented fragments (in line with a 60% average margin in single-member districts; the actual US average varies between 60-70%), then 18% would be unrepresented nationwide. If the unrepresented fragment is 1/2 the size of the represented one, 16% would be unrepresented nationwide.)

.........

Here's the current version of the electorama article on PAL representation:

PAL (Proportional, Accountable, Local) representation is a system for electing a proportionally-representative legislature. It's designed to be a gentle change from a single-member-district system; districts can remain unchanged, and if single-member districts are giving fair proportions from cohesive parties, PAL representation will elect exactly the same members. The difference is that most representatives will represent multiple districts, and each district will have multiple representatives (one from each winning party). This allows each voter to know who their representative is, while preserving ballot secrecy. Thus, whereas currently only 60-70% of US voters voted for their representative, and many of those because they have no real choice, with PAL voting over 80% overall, and over 95% in large states, would be guaranteed have a representative whom they'd supported directly or indirectly.

The basic idea is:

  • Candidates pre-announce their rank-ordering of the parties (starting with their own party) and may optionally approve/disapprove within each party candidates. Their votes will never be transferred to disapproved candidates.
  • Voters may vote on the candidates in their or nearby districts, or write in candidates from farther off. Votes are delegated by default but optionally, voters may refuse to delegate or vote approval-style.
  • Each delegated ballot is transformed into the pre-announced vote of the candidate it supports.
  • A legislature is elected by a version of STV (with fractional transfers and a Droop quota.)
  • Each district "drafts" one member of each elected party from the elected slate.
  • Your representative is the member of the party you voted for who is representing your district.

Contents

 [hide]

[edit]Full Procedure

  • Candidates pre-announce their rank-ordering of the parties (starting with their own party) and may optionally disapprove of any other candidates.
  • Voters may vote on the candidates in their or nearby districts, or write in candidates from farther off.

First, to simplify the ballots, the population is separated into a "district" for each seat, and "districts" are grouped into sets of 2 or 3 "co-districts". The ballot for each district lists the incumbents and candidates from that district in a larger font, and the candidates from its co-districts below that in a smaller font. Write-ins may be used to vote for candidates from other districts not listed on the ballot, so the districts only matter for ballot simplicity (Voters do not want to have a ballot with many dozens of candidates on it, but write-ins allow full freedom for those voters who want it). Larger parties will usually run one candidate per district; smaller parties may just run one candidate per co-district set.

  • Each vote is transformed into the pre-announced party preference order and individual approvals/disapprovals of the candidate it chooses.
  • A legislature is elected by a modified version of STV, using the following steps:
  1. The iteration number I is set to 1. (Higher iteration numbers mean lower quotas. The process will be run from scratch with an increasing iteration number until a full slate of candidates reaches the quota)
  2. The quota Q is set to the Droop Quota; that is, to the rational number (V+1)/(S+I), where V is valid votes and S is seats.
  3. Delegated votes first count full-weight for their chosen candidate. Once that candidate is elected or eliminated, a vote is divided equally among all non-disapproved, non-eliminated members of the top party remaining on that ballot with any such members.
  4. Undelegated votes are divided equally among all approved, non-eliminated candidates on that ballot.
  5. Any candidates who reach the quota are immediately and simultaneously elected, and their ballots are reweighted to eliminate a Droop quota.
  6. If there are no candidates who reach the quota, the party with the fewest votes is identified, and the candidate from that party with the fewest votes is eliminated. All votes for that candidate are reassigned as outlined above.
  7. If the above finishes without electing a full slate, the process starts from scratch with a lower quota:
    • All ballots are reweighted to 1
    • All elected representatives return to being hopeful candidates
    • The iteration number is increased by 1. This reduces the quota Q, as if it were the Droop quota for a legislature one seat bigger.
    • The counting process is rerun from scratch, starting with step 3.
  8. Each district "drafts" one member of each elected party from the elected slate. The draft proceeds as follows:
    • First, each representative is drafted by their home district.
    • From then on, the draft proceeds in descending order of votes. That is, if more votes from district 1 go to candidate A than any other eligible district:candidate pair, then A is drafted to that district. Eligibility rules are as follows:
    • All representatives from a party must be drafted N times before any representative from that party may be drafted N+1 times.
    • No district may draft two representatives from the same party.

Your representative is the member of the party you voted for who is representing your district. If no member of the party you voted for was elected, then you may look at the public ballot of your chosen candidate to see which of your district's representatives is yours.

[edit]Optional party threshold

Optionally, one additional rule can be added to modify step 5 above:

  • No representative may be elected unless their party got at least T votes, where T is some party threshold.

This would encourage small parties to join into coalitions, and thus promote a less-fragmented legislature. There are various options for T. It could be as high as 5%, similar to the 5% threshold used in the German parliament. Or it could be as low as V/(S+I-1) (that is, V/S, if the process completes in just one iteration); this would actually allow independent candidates to be their own "party", but only if they have enough support to fully deserve one of the S seats.

This rule complicates the system somewhat, so it is not recommended if the PAL representation is to be implemented by a voter referendum. If the system is being passed by a legislature, they may be more concerned about fragmentation, so they could use a relatively-high 5% threshold. And if the system is implemented by a constitutional convention, a V/(S+I-1) threshold is ideally fair.

[edit]Sample Ballot

District 5 ballot
Vote for one candidate to delegate your vote, or otherwise vote for as many candidates as you approve:
▢ John Adams (Yellow Party, district 5) (Prefers Yellow Party except for Zapatero)
▢ Michelle Bachelet (Silver Party, district 5) (Prefers Silver Party, then Purple Party)
▢ Winston Churchill (Purple Party, district 5) (Prefers Purple Party, then Silver Party except for Yarrow)
▢ Alfred Deakin (Yellow Party, district 6) (Prefers Yellow Party then Silver party except for Yarrow)
▢ John Edwards (Silver Party, district 6) (Prefers Silver Party then Purple party)
▢ Vicente Fox (Orange Party, district 6) (Prefers Orange Party then Silver party)
▢ Inder Gujral (Yellow Party, district 4) (Prefers Yellow Party)
▢ Stephen Harper (Silver Party, district 4) (Prefers Silver Party)

▢ ________________________(write-in)
▢ ________________________(write-in)

If you only vote for one candidate who does not win, your vote may help elect that candidate's preferences, unless you check the box below:
▢ Do not delegate.

Note: When you vote for just one candidate, you are delegating your vote to him or her. This allows your vote to help elect a similar candidate if your favorite cannot win. Your vote will never go to a candidate who is not in your favorite's declared preferences listed above.

If you vote for more than one candidate, or if you vote for a candidate with no declared preferences, your vote is not delegated. In that case, it does not matter whether you check the "do not delegate" box above.

[edit]Example

Tennessee's four cities are spread throughout the state

Imagine that Tennessee is having an election on where to locate 3 public universities. The population of Tennessee is concentrated around its four major cities, which are spread throughout the state. For this example, suppose that the entire electorate lives in these four cities, and that everyone wants to live as near as many universities as possible.

The candidate sites for the university are:

  • Memphis on Wikipedia, the state's largest city, with 42% of the voters, but located far from the other cities
  • Site 1
  • Site 2
  • Nashville on Wikipedia, with 26% of the voters, near the center of Tennessee
  • The "Eastern Party", composed of:

The preferences of the voters would be divided like this:

42% of voters
(close to Memphis)
26% of voters
(close to Nashville)
15% of voters
(close to Chattanooga)
17% of voters
(close to Knoxville)
  1. Memphis 1
  2. Rest of Memphis party:
    • Memphis 2
  3. Nashville
  4. Eastern party
    • Chattanooga
    • (Knoxville not approved)
  1. Nashville
  2. Eastern party:
    • Chattanooga
    • Knoxville
  3. Memphis party
  1. Chattanooga
  2. Rest of Eastern Party:
    • Knoxville
  3. Nashville
  4. Memphis party
  1. Knoxville
  2. Rest of Eastern Party:
    • Chattanooga
  3. Nashville
  4. Memphis party


The quota is (100.0008%/(3+1))=25.0002% (The small fraction represents one extra virtual voter, to ensure that the quota cannot be met by four different sites). Since both Memphis (site 1) and Nashville are over the quota, both are elected first. Memphis votes are multiplied by 17/42 and transferred to Memphis site 2, and Nashville votes are multiplied by 1/26 and then split evenly between Chatanooga and Knoxville. Totals are now:

  • (Memphis 1: 25.0002% (elected))
  • Memphis 2: ~17% (actually 16.9998%)
  • (Nashville: 25.0002% (elected))
  • Chatanooga: ~15.5% (actually, 15.4999%)
  • Knoxville: ~17.5% (actually 17.4999%)

The party with the fewest remaining votes is the Memphis party. Within that party, Memphis 2 is the site with the fewest votes (in fact, the only remaining site), so even though it has more votes than Chatanooga, Memphis 2 is eliminated. The votes are pass over the already-elected Nashville to tranfer to the Eastern party. Within that party, Memphis disapproved Knoxville, so the full total is transferred to Chatanooga. Chatanooga now has ~32.5%, more than the 25% quota, so it is the third and final site.

If Knoxville had not joined a party with Chatanooga, then Chatanooga would have been eliminated, and Knoxville would have been the final site. But Chatanooga could have responded by threatening to prefer a second Nashville site, or even Memphis 2, over Knoxville, if Knoxville would not cooperate in the Eastern party. In the end, Knoxville's strategy may or may not have worked. In general, such strategic gamesmanship would be less profitable and more dangerous in a real election, with more seats overall as well as a significant degree of polling uncertainty.

[edit]Advantages

[edit]P

  • Proportional
  • Thus, a large majority of voters have real representation
  • Each representative is elected with the same number of votes.
  • Prudent; not a radical change from single-member districts
  • No redistricting necessary
  • If:
  • all votes are for one of the two main-party candidates in the voter's district,
  • all candidates approve everyone from their party
  • and the districts are divided fairly so that plurality would give a proportional result
... then PAL representation (like Balinski's "Fair Representation") gives the same results as plurality. These assumptions will not generally be perfectly true, but they will generally be close to true, so PAL representation will give results that are recognizably similar to those of single-member districts. It is hoped that this would make it a more acceptable system to politicians who have won under single-winner rules.

[edit]A

  • Accountable
  • Voters, not party bureaucrats, decide which members of a given party get seated.
  • Since the total votes needed for election is higher, the "margin of victory" is reduced. There are no safe, gerrymandered seats where corrupt representatives can hide.

[edit]L

  • Local
  • Representatives know who is a constituent and voters know who is their representative.
  • Neighbors can organize to lobby their shared representatives.
  • Fair attention for local issues.

[edit]Justification

PAL representation is inspired by Michel Balinski's "Fair Representation" and by SODA voting. From the former, which is used for municipal elections in Belgium, it inherits the combination of geographical districts and proportionality. However, unlike Fair Representation, each candidate elected by PAL representation has received (directly or indirectly) the same number of votes. From SODA voting, PAL representation inherits the simple, spoilproof ballot format and the optional vote delegation.

A modified version of STV is used as the proportional system for simplicity. Other proportional systems might also work (although a non-LNH system might put perverse incentives on candidates). The equal ranking, and resulting fractional division of votes, is necessary for three reasons. First, it allows for approval-style votes to be counted without complicating the ballot. Second, it allows candidates to exercise judgment independently from their party (disapproving of certain party members), but keeps the voter's judgment as primary. If candidates couldn't exercise judgment, parties would have to waste energy keeping out "crazy" candidates who affiliate only because of the transfer votes they might get. If candidates could fully-rank within the party, as would happen if the PR system were standard STV, there would be too many opportunities for logrolling, at a level of detail where voters wouldn't realistically keep track or hold candidates accountable. Third, equal-ranking allows us to claim that this system could, under reasonable circumstances, elect exactly the same representatives as a non-gerrymandered single-member-district system.

Jameson Quinn

unread,
Oct 25, 2011, 9:36:34 AM10/25/11
to EM, electionsciencefoundation
OK, I'm turning into a bit of a spammer on this issue, but...

Here's some discussion of PAL representation's legality in the US. A state could pass a law to use FMV (a true proportional system) until the anti-PR federal law is repealed, then automatically switch to PAL representation (an easy and smooth transition; FMV was a major inspiration for PAL). Basically, I think it's an important fact that FMV is the only known proportional system compatible with current US federal law. (There are no US constitutional barriers to PR, just federal law).

Jameson Quinn

Andy Jennings

unread,
Oct 25, 2011, 1:45:51 PM10/25/11
to Jameson Quinn, EM, electionsciencefoundation
Hi Jameson,

I looked over it.  I didn't see any technical problems immediately, but I'm going to try to re-read it a few more times and keep thinking it over.  My emotional response, though, is that it's probably beyond the complexity limit for actual implementation anytime soon.  The idea that "candidates from your district are in a bigger font and candidates outside your co-districts aren't listed at all but you can write them in" is a clever trick, but I don't know if people will go for it.

Clarification on terminology: When you say "Fair Representation", is that the same thing as "Fair Majority Voting"?
- I know "Fair Majority Voting" from here: http://mathaware.org/mam/08/EliminateGerrymandering.pdf
- I know Balinski has a more complicated system that allows multiple winners per district.  I forget what it's called, biproportional apportionment, maybe?
- When I hear "Fair Representation", I think of his book with Young (http://books.google.com/books?id=eBFNzSeAv_sC) which is all about congressional apportionment, not about PR or voting systems.

More general questions:

What are your main priorities here?  Just to find some way to use the SODA ballot for a PR election?  Do you see locality as a must-have or as a nice-to-have that also decreases the complexity of the ballot?

What advantages does this have over straight FMV?  Just that it uses a SODA ballot?  The fact that you are almost certain to have a representative of your own party?

Why not just use straight FMV with a "districting" step for each party afterwards?  For example, say we use FMV and party A gets 5 seats, party B gets 4 seats, and party C gets 1 seat.  We take a clean map and divide it into five regions so that there are an equal number of A voters in each region.  Those are the districts for party A voters.  Take another clean map and divide it into four regions so there are an equal number of B voters in each region, and those are the districts for the party B voters.  The party C map doesn't get divided up because they all are represented by the one seated C candidate.

Maybe you could use some kind of weighted Vonoroi to choose the districts so that the districts cluster around the homes of the seated representatives from that party.  Or maybe you would use splitline or some other districting method.  Or maybe you let each party decide how to divide up their map themselves.  I'm hoping there wouldn't be much gerrymandering because these "party districts" don't affect future elections in any way.  They're just done so you can show people "if you voted for party A, then look at this map to see who your party A representative is."

Or maybe you don't care that the number of A voters is exactly the same in each of party A's regions, then you can just agglomerate the legislative districts into relatively balanced co-districts.

Andy



----
Election-Methods mailing list - see http://electorama.com/em for list info


Bruce Gilson

unread,
Oct 25, 2011, 1:52:31 PM10/25/11
to electio...@googlegroups.com
On Tue, Oct 25, 2011 at 8:32 AM, Jameson Quinn <jameso...@gmail.com> wrote:
Yesterday I posted two messages to the Election Methods list about PAL representation, a new PR system designed as a replacement for single-member districts. The system combines elements from STV, Balinski's "Fair Representation", and SODA. 

Basically nothing in the system as I described it was in itself a new idea. My goal was not to create an innovative system, but to blend existing ideas to find a proportional system that would be acceptable to all existing interests, including voters, theorists/activists like us, incumbent politicians, and parties.

I believe that a system like this is the only kind of PR that could pass in a US environment, and also well-suited to the UK. I'm a little discouraged that a development I regard as truly important got no comment from the list.
 
The list you posted it on is not one I get in my e-mail feed. Thus I had no knowledge of it.
 
While I'm not sure that your proposed system is "the only kind of PR that could pass in a US environment," given that New Zealand adopted a German-style MMP system, I think it looks promising from a first cursory examination. I really think I need to look at it in more detail, but given my own situation (a moderate Republican in the heart of far-left Democratic territory), it looks like almost the only way I've seen that could give me something like "my" representative, which alone makes it an attractive system. But I will get back to you later after I've absorbed it more thoroughly.

Bruce Gilson

unread,
Oct 25, 2011, 1:57:11 PM10/25/11
to electio...@googlegroups.com


On Tue, Oct 25, 2011 at 1:45 PM, Andy Jennings <elec...@jenningsstory.com> wrote:
[...]
Why not just use straight FMV with a "districting" step for each party afterwards?  For example, say we use FMV and party A gets 5 seats, party B gets 4 seats, and party C gets 1 seat.  We take a clean map and divide it into five regions so that there are an equal number of A voters in each region.  Those are the districts for party A voters.  Take another clean map and divide it into four regions so there are an equal number of B voters in each region, and those are the districts for the party B voters.  The party C map doesn't get divided up because they all are represented by the one seated C candidate.

Maybe you could use some kind of weighted Vonoroi to choose the districts so that the districts cluster around the homes of the seated representatives from that party.  Or maybe you would use splitline or some other districting method.  Or maybe you let each party decide how to divide up their map themselves.  I'm hoping there wouldn't be much gerrymandering because these "party districts" don't affect future elections in any way.  They're just done so you can show people "if you voted for party A, then look at this map to see who your party A representative is."
 
This is another nice idea. It is, as you say, simpler than Jameson's proposal, and thus perhaps easier to explain to the people who might adopt it. Personally, either Jameson's plan or yours would meet with my greatest need -- having a representative I could call "my own" who is not ddiametrically opposed to everything I want. So I could accept either plan.

Jameson Quinn

unread,
Oct 25, 2011, 2:56:07 PM10/25/11
to Andy Jennings, EM, electionsciencefoundation


2011/10/25 Andy Jennings <elec...@jenningsstory.com>

Hi Jameson,

I looked over it.  I didn't see any technical problems immediately, but I'm going to try to re-read it a few more times and keep thinking it over.  My emotional response, though, is that it's probably beyond the complexity limit for actual implementation anytime soon.  The idea that "candidates from your district are in a bigger font and candidates outside your co-districts aren't listed at all but you can write them in" is a clever trick, but I don't know if people will go for it.

(Note: I've renamed "co-district" as "super-district")

The super-district idea and ballot design are not fundamental. For simplicity, it would work fine if only the candidates from your local district were available, as long as you could write-in candidates from other districts.


Clarification on terminology: When you say "Fair Representation", is that the same thing as "Fair Majority Voting"?

Yes. Oops. I'd already fixed this error on the page.
 
- I know "Fair Majority Voting" from here: http://mathaware.org/mam/08/EliminateGerrymandering.pdf
- I know Balinski has a more complicated system that allows multiple winners per district.  I forget what it's called, biproportional apportionment, maybe?

Don't know that one. Can you find a link?
 
- When I hear "Fair Representation", I think of his book with Young (http://books.google.com/books?id=eBFNzSeAv_sC) which is all about congressional apportionment, not about PR or voting systems.

Sorry, mistake.

 

More general questions:

What are your main priorities here?  Just to find some way to use the SODA ballot for a PR election?  Do you see locality as a must-have or as a nice-to-have that also decreases the complexity of the ballot?

The main priority is to make a proportional system which can use the existing district structure, but doesn't suffer from FMV's flaws (see next section below)
 

What advantages does this have over straight FMV?  Just that it uses a SODA ballot?  The fact that you are almost certain to have a representative of your own party?

Compared to straight FMV, this has 4 advantages. In descending importance:
1. In FMV, a minority candidate can beat a majority candidate in that same district. In PAL, there is a "chain of custody" of the delegated votes, so you can see that each winner has attained exactly the same quota of votes.
2. FMV gives too much power to the parties. A vote for one candidate from party X is assumed to be equivalent to voting for any other candidate from that party. In PAL, both candidates and voters have some recourse to be able to vote for party X but not candidate Y in that party. (Voter recourse is the purpose of using the SODA ballot. If the SODA ballot were considered too complex, I'd consider just write-ins to be an acceptable, though inferior, substitute; because with full choice, you could probably find a candidate whose preferences were close enough to yours.)
3. As you say, PAL lets you track which representative from your party is "yours".
4. PAL lets votes for unsuccessful parties be reassigned to larger parties.

I consider 1 and 2 above to be important enough to design a new system. Yes, 1 is only symbolic, really; but it's a doozy. I can't imagine convincing a majority to support FMV in spite of this simple opposition talking point. Also, note that fixing 2 makes the fix for 1 more than just symbolic.

Why not just use straight FMV with a "districting" step for each party afterwards?  For example, say we use FMV and party A gets 5 seats, party B gets 4 seats, and party C gets 1 seat.  We take a clean map and divide it into five regions so that there are an equal number of A voters in each region.  Those are the districts for party A voters.  Take another clean map and divide it into four regions so there are an equal number of B voters in each region, and those are the districts for the party B voters.  The party C map doesn't get divided up because they all are represented by the one seated C candidate.

That would work to fix issue 3 above, but it seems to me more complicated than my system. It is also a more radical change, especially from the perspective of an incumbent. Also, it's more unlike FMV, which is the only PR system legal under current federal law. 
 

Maybe you could use some kind of weighted Vonoroi to choose the districts so that the districts cluster around the homes of the seated representatives from that party.  Or maybe you would use splitline or some other districting method.  Or maybe you let each party decide how to divide up their map themselves.  I'm hoping there wouldn't be much gerrymandering because these "party districts" don't affect future elections in any way.  They're just done so you can show people "if you voted for party A, then look at this map to see who your party A representative is." 

This would work. Again, it only fixes issue 3.
 

Or maybe you don't care that the number of A voters is exactly the same in each of party A's regions, then you can just agglomerate the legislative districts into relatively balanced co-districts.

That's basically what PAL does.

Jameson Quinn

unread,
Oct 25, 2011, 3:00:04 PM10/25/11
to electio...@googlegroups.com


2011/10/25 Bruce Gilson <brg...@gmail.com>

While I'm not sure that your proposed system is "the only kind of PR that could pass in a US environment,"

I didn't say this exact system, I said something "like" it. In other words, I think that any system which doesn't leave existing representatives with some meaningful "home base" will be effectively vetoed by them.

given that New Zealand adopted a German-style MMP system,

It's a good point, but New Zealand is smaller than the US, with fewer veto points. It's also not a federal system.

Jameson

Andy Jennings

unread,
Oct 25, 2011, 3:31:16 PM10/25/11
to Jameson Quinn, EM, electionsciencefoundation
On Tue, Oct 25, 2011 at 11:56 AM, Jameson Quinn <jameso...@gmail.com> wrote:


2011/10/25 Andy Jennings <elec...@jenningsstory.com>
Hi Jameson,

I looked over it.  I didn't see any technical problems immediately, but I'm going to try to re-read it a few more times and keep thinking it over.  My emotional response, though, is that it's probably beyond the complexity limit for actual implementation anytime soon.  The idea that "candidates from your district are in a bigger font and candidates outside your co-districts aren't listed at all but you can write them in" is a clever trick, but I don't know if people will go for it.

(Note: I've renamed "co-district" as "super-district")

The super-district idea and ballot design are not fundamental. For simplicity, it would work fine if only the candidates from your local district were available, as long as you could write-in candidates from other districts.


Clarification on terminology: When you say "Fair Representation", is that the same thing as "Fair Majority Voting"?

Yes. Oops. I'd already fixed this error on the page.
 
- I know "Fair Majority Voting" from here: http://mathaware.org/mam/08/EliminateGerrymandering.pdf
- I know Balinski has a more complicated system that allows multiple winners per district.  I forget what it's called, biproportional apportionment, maybe?

Don't know that one. Can you find a link?

Here are some references:

Michel Balinski, “Apportionment : uni- and bi-dimensional,” in B. Simeone et F. Pukelsheim (Eds.), Mathematics 
and Democracy. Recent Advances in Voting Systems and Collective Choice, Springer, Berlin and Heidelberg, 2006, 
43-53.

Michel Balinski and Friedrich Pukelsheim, “Die Mathematik der doppelten Gerechtigkeit,” Spektrum der 
Wissenschaft, April 2007, 76-80.

Michel Balinski and Friedrich Pukelsheim, “Matrices and politics,” in E. P. Liski, J. Isotalo, S. Puntanen and G. P.
H. Styan (Eds.), Festschrift for Tarmo Pukkila, Department of Mathematics, Statistics and Philosophy, University of 
Tampere, Finland, 2006, 233-242.

Jim Riley

unread,
Oct 26, 2011, 7:11:43 AM10/26/11
to electio...@googlegroups.com, EM
I think that is an unrealistic interpretation of current law in the United States.

There has been a requirement that representatives be elected from districts since at least 1872.  That was the first reapportionment after the Civil War, and the northern hegemons were taking advantage of the lack of southerners and Democrats to impose their standards upon all the States.  The reapportionment bill not only required election from districts, but also that districts be defined prior to the next election.  It also set the uniform congressional election date (1st Tuesday after 1st Monday in November) and required elections to be conducted by paper ballot.

It was later discovered that some northern States didn't like being told what to do, and the standards were relaxed.  It wasn't until 1960 that all States held their elections in November.  And the districting standard was somewhat relaxed.  If a State received additional representatives, they could be elected at large until they redistricted (and in some cases this was not accomplished during an entire decade), and if they lost representatives they could elect all representatives at large until they districted (and some States went a decade like that as well).  And states that elected all representatives at large could continue that practice whenever they lost or gained representatives.

The intent of the change in the 1960s was simply to require immediate redistricting.  The reason the change was made as an amendment to a private bill was because the original bill got bogged down in procedure, and the amendment got stuck in a bill that was going to pass.

Current law requires representatives to be "elected by district", not that each district "have" a representative.   If a candidate receives fewer votes from voters in a district, can he be said to have been elected by that district?

But let's imagine that you convinced a State that Congress left a loophole.  You would likely then have a legislature where the party that is advantaged by district elections, is also advantaged in legislative elections.  So you go to Maryland or Illinois and tell the legislature that if they adopted this new system, that more Republicans would be elected to Congress.  That their good friends who they had just contributed the maximum amount to, might lose his seat, and all their efforts at gerrymandering were for naught.   And in particular, you tell them that it is the white Democrats from swingy districts that will lose their seats.  If they were honest, they would tell you it was not to their partisan advantage.  But they could simply say that it doesn't appear to be legal and thank you for your time.

But even if you got a legislature to pass the bill, you are likely to go to court, which would be skeptical at best.  If you were lucky, you might convince them that the decision is for Congress to make.  They passed the law, and they have sole authority to judge election contests.  They will interpret what the law means, and are unlikely to interpret it in a way where a "losing" candidate is elected.

The Balinski paper has some errors.  He mixes up "census tract" and "census block".

He also got confused on the court decisions.  Pennsylvania passed a redistricting plan where they got careless with population equality, and had some odd little appendages and ended up in the Supreme Court.  Justice Kennedy thought he had his judiciable standard for a "partisan gerrymander".  The court remanded the case back to the district court.  Meanwhile Pennsylvania had fixed up their map so that it was perfectly equal (+/- 1) and that was upheld by the federal courts.

The Texas case had been appealed up to the Supreme Court, and it sat around for several months while the court considered whether to take the case.  Finally, the court remanded the case back to the district court, to consider whether the Pennsylvania case applied to Texas (the ruling on the Pennsylvania case occurred after the district court had ruled in Texas).   This got the plaintiffs all hopeful, so they started making all kinds of new arguments in hope that the Supreme Court would bite.  One was that Texas could not use census numbers for redistricting because everyone knew that all this nonsense about equal population +/-1 was a legal fiction.  The district court and appeals court determined that the Pennsylvania case did not apply.  Eventually (in 2006), the Supreme Court did overturn part of the Texas plan, by inventing a new rule, but not because it was a political gerrymander.

Jameson Quinn

unread,
Oct 26, 2011, 7:24:33 AM10/26/11
to Andy Jennings, EM, electionsciencefoundation


2011/10/25 Andy Jennings <elec...@jenningsstory.com>



On Tue, Oct 25, 2011 at 11:56 AM, Jameson Quinn <jameso...@gmail.com> wrote:


2011/10/25 Andy Jennings <elec...@jenningsstory.com>
Hi Jameson,

I looked over it.  I didn't see any technical problems immediately, but I'm going to try to re-read it a few more times and keep thinking it over.  My emotional response, though, is that it's probably beyond the complexity limit for actual implementation anytime soon.  The idea that "candidates from your district are in a bigger font and candidates outside your co-districts aren't listed at all but you can write them in" is a clever trick, but I don't know if people will go for it.

(Note: I've renamed "co-district" as "super-district")

The super-district idea and ballot design are not fundamental. For simplicity, it would work fine if only the candidates from your local district were available, as long as you could write-in candidates from other districts.


Clarification on terminology: When you say "Fair Representation", is that the same thing as "Fair Majority Voting"?

Yes. Oops. I'd already fixed this error on the page.
 
- I know "Fair Majority Voting" from here: http://mathaware.org/mam/08/EliminateGerrymandering.pdf
- I know Balinski has a more complicated system that allows multiple winners per district.  I forget what it's called, biproportional apportionment, maybe?

Don't know that one. Can you find a link?

Here are some references:

Michel Balinski, “Apportionment : uni- and bi-dimensional,” in B. Simeone et F. Pukelsheim (Eds.), Mathematics 
and Democracy. Recent Advances in Voting Systems and Collective Choice, Springer, Berlin and Heidelberg, 2006, 
43-53.

At a quick read, this is basically just a prettier formalism for FMV. Yes, it is good math; no, it is not a politically- or legally-significant change. 

Oh, wait... it also allows multi-member districts. But not post-hoc overlapping districts as in PAL.

Michel Balinski and Friedrich Pukelsheim, “Die Mathematik der doppelten Gerechtigkeit,” Spektrum der 
Wissenschaft, April 2007, 76-80.

I can't read German. 

Michel Balinski and Friedrich Pukelsheim, “Matrices and politics,” in E. P. Liski, J. Isotalo, S. Puntanen and G. P.
H. Styan (Eds.), Festschrift for Tarmo Pukkila, Department of Mathematics, Statistics and Philosophy, University of 
Tampere, Finland, 2006, 233-242.

This appears to use the prettier formalism to discuss a multi-member-district FMV-like system in Zurich.

See also: S. Maier, P. Zachariassen, and M. Zachariasen, “Divisor-based biproportional apportionment in electoral systems: A real-life benchmark study,” Management Science 56, no. 2 (2010): 373–387.

This looks at algorithm performance and result quality (over several measures) in the general problem of multi-member districts and global proportionality, for various sources of real-world and simulated (including deliberately pathological) data.

All of the above have the same problem I see with FMV: in the single-member-district case, they could elect a minority candidate over the majority candidate for that district. PAL resolves this problem by making vote transfers explicit and, from the voter's point of view, optional and thus intentional. Thus PAL elects winners as if it were a "single-member biproportional" system, but it provides a much clearer rationale for the "discordant" district-level results: the loser may have gotten more direct votes, but the winner got a higher direct+transferred total; and all winners reached the same overall quota.

This is the main idea of PAL.

JQ

Jameson Quinn

unread,
Oct 26, 2011, 7:45:10 AM10/26/11
to electio...@googlegroups.com, EM


2011/10/26 Jim Riley <jim...@pipeline.com>

I think that is an unrealistic interpretation of current law in the United States.

Thanks a lot for your legal analysis; you clearly know a lot more about this than I. I won't address your (very interesting) case-by-case discussion, but I will respond to your point about the politics of this:
 
But let's imagine that you convinced a State that Congress left a loophole.  You would likely then have a legislature where the party that is advantaged by district elections, is also advantaged in legislative elections.  So you go to Maryland or Illinois and tell the legislature that if they adopted this new system, that more Republicans would be elected to Congress. 

Very good point. A state legislature which performs a partisan gerrymander, will not want to undermine all that hard work by adopting FMV or PAL...
 
That their good friends who they had just contributed the maximum amount to, might lose his seat, and all their efforts at gerrymandering were for naught.   And in particular, you tell them that it is the white Democrats from swingy districts that will lose their seats.  If they were honest, they would tell you it was not to their partisan advantage.  But they could simply say that it doesn't appear to be legal and thank you for your time.

 ...and thus they will tend to (or pretend to) interpret any gray area of the law against the new system.

This is a pretty solid argument that FMV will not pass until the 1967 law is overturned.

It is not an argument that the law will be impossible to overturn. Undermining partisan gerrymanders on a state-by-state level will tend to balance out more on a national level. And counting all the states with under 3 reps plus all the states with reasonably non-partisan redistricting rules (AK, AZ, CO, CA, HI, ID, IA, MO, MT, NJ, PA, WA), you can reach a majority of the House; so gerrymandered representatives don't have veto power over an overturn.

Once the law is overturned, the fight will move to the states. There, it will be a tough fight, but there are two ways to overcome the pro-gerrymander bias Jim described. First, voter initiatives (in states which allow them); and second, states where partisan control has switched since the last redistricting. 

Jameson

Warren D. Smith (CRV cofounder, http://RangeVoting.org)

unread,
Oct 26, 2011, 6:50:31 PM10/26/11
to The Center for Election Science

ok, trying to comprehend the PAL system.
Random comments/questions follow.

>A closed list system would be (rightly) attacked as a power grab by party bureaucrats.

--err, it would be better to rewrite this as
'A closed list system would be (rightly) attacked as "a power grab by
party bureaucrats".'

>A mixed member system is an ugly hybrid. To me, US democratic ideals are too egalitarian to accept that there could be two different kinds of representative.

--a rather unconvincing argument. US has way more "egalitarian
ideals" than Germany? Who knew.

>More seriously, a mixed member system would be totally unacceptable to
existing incumbents, as it would draw too many of them out of their
existing
districts.

--also [relatedly] highly unconvincing. I think many politicians
don't give a damn that they are "drawn out of" their districts. The
USA has in fact had in the past, states with "at large" congressmen,
so we know for sure you are wrong.

>The basic idea is:
> - Candidates pre-announce their rank-ordering of the parties (starting
with their own party) and may optionally approve/disapprove within
each
party candidates.

--I cannot parse the above sentence after the word "optionally."

>Their votes will never be transferred to disapproved
candidates.
> - Voters may vote on the candidates in their or nearby districts, or
write in candidates from farther off.

--you neglected to explain the minor matter of what a "vote" IS!!!!
Hello?

>Votes are delegated by default but
optionally, voters may refuse to delegate or vote approval-style.

--as opposed to what? (And the wording "may refuse to X or Y"
has unclear meaning. I in fact do not know which of the two meanings
you intended,
and I also do not know what happens when they refuse, I presume you
have some alternative in mind.)

> - Each delegated ballot is transformed into the pre-announced vote of the
candidate it supports.

> - A legislature is elected by a version of
STV<http://wiki.electorama.com/wiki/STV> (with
fractional transfers and a Droop quota.)

--so the votes were rank-order ballots? (implied?) THis
unfortunatley cuases
things to be complicated, since Droop STV is complicated. Which is
fine, but if your goal was to make things simple, you just blew it.

> - Each district "drafts" one member of each elected party from the
elected slate.

--which one?

> - Your representative is the member of the party you voted for who is
representing your district.

--so voters vote for parties not people???

--well, I quit reading here. Maybe you've got a nice idea (a "big
deal" as you put it),
but up to this point it has been written incomprehensibly.

Warren Smith

unread,
Oct 26, 2011, 7:10:37 PM10/26/11
to The Center for Election Science
> - Your representative is the member of the party you voted for who is
representing your district.

--in order for this to be true, your rep needs to know, i.e. it needs
to be public information that,
you voted for his party.

So you've just abandoned ballot secrecy.

Now if (as in the whole first part of your post) you're looking for something
that would be unacceptable in the USA... you just found it.

Jameson Quinn

unread,
Oct 26, 2011, 7:32:32 PM10/26/11
to electio...@googlegroups.com
You know who your representative is, but they don't. All they know is that you're in one of their districts.

So if you wrote a letter to a Democratic representative asking for right-to-work legislation, or to a Republican representative asking for a millionaire surtax, and claimed that they were your representative... they might be a little suspicious. Vice versa, and they'd probably believe you.

And in the end, would it be so bad if you had recourse to the constituent services of 2 or 3 representatives, instead of just one? Yes, there is some inefficient overlap and some dilution; but that would open up room for specialization, too. "Rep X has better attention for family immigration issues, rep Y for ..."

JQ

2011/10/26 Warren Smith <warre...@gmail.com>

Jameson Quinn

unread,
Oct 26, 2011, 7:48:19 PM10/26/11
to electio...@googlegroups.com
>More seriously, a mixed member system would be totally unacceptable to
existing incumbents, as it would draw too many of them out of their
existing
districts.

--also [relatedly] highly unconvincing.  I think many politicians
don't give a damn that they are "drawn out of" their districts.  The
USA has in fact had in the past, states with "at large" congressmen,
so we know for sure you are wrong.

The point is that incumbents like the status quo. Obviously, going to PR means changing that status quo, but it is a reasonable goal to do so as little as possible.
 

>The basic idea is:
>   - Candidates pre-announce their rank-ordering of the parties (starting
  with their own party) and may optionally approve/disapprove within
each
  party candidates.

--I cannot parse the above sentence after the word "optionally."

They say for example "I like Red (candidates A and C but not B and D); then Green (all of them); then Blue (only candidate X)".
 

>Their votes will never be transferred to disapproved
  candidates.
>   - Voters may vote on the candidates in their or nearby districts, or
  write in candidates from farther off.

--you neglected to explain the minor matter of what a "vote" IS!!!!
Hello?

A vote is either a delegated bullet vote or an undelegated approval vote.
 

>Votes are delegated by default but
  optionally, voters may refuse to delegate or vote approval-style.

--as opposed to what?  (And the wording "may refuse to X or Y"
has unclear meaning.  I in fact do not know which of the two meanings
you intended,
and I also do not know what happens when they refuse, I presume you
have some alternative in mind.)

See the STV procedure for the meanings. 

>   - Each delegated ballot is transformed into the pre-announced vote of the
  candidate it supports.

>   - A legislature is elected by a version of
STV<http://wiki.electorama.com/wiki/STV> (with
  fractional transfers and a Droop quota.)

--so the votes were rank-order ballots?  (implied?)  THis
unfortunatley cuases
things to be complicated, since Droop STV is complicated.  Which is
fine, but if your goal was to make things simple, you just blew it.

Simple for voters != simple to fully specify. The former is more important.

I am not wedded to using STV here. If there's another proportional system which works, I'd be open to using it.
 

>   - Each district "drafts" one member of each elected party from the
  elected slate.

--which one?

See the section just below. This is just the outline, not the full rules.
 

>   - Your representative is the member of the party you voted for who is
  representing your district.

--so voters vote for parties not people???

You vote for a person, that person has a party. 

--well, I quit reading here.  Maybe you've got a nice idea (a "big
deal" as you put it),
but up to this point it has been written incomprehensibly.

It's been edited considerably. I'll revise it with this email in mind tomorrow morning.

JQ

Warren D. Smith (CRV cofounder, http://RangeVoting.org)

unread,
Oct 28, 2011, 1:54:33 PM10/28/11
to The Center for Election Science


> > >The basic idea is:
> > >   - Candidates pre-announce their rank-ordering of the parties (starting
> >    with their own party) and may optionally approve/disapprove within
> > each
> >   party candidates.
>
> > --I cannot parse the above sentence after the word "optionally."
>
> They say for example "I like Red (candidates A and C but not B and D); then
> Green (all of them); then Blue (only candidate X)".

--aha. Quite complicated.

> A vote is either a delegated bullet vote or an undelegated approval vote.

--aha.

> > >   - Each district "drafts" one member of each elected party from the
> >   elected slate.
> > --which one?
> See the section just below. This is just the outline, not the full rules.

> > >   - Your representative is the member of the party you voted for who is
> >   representing your district.
> > --so voters vote for parties not people???
> You vote for a person, that person has a party.

--so your system screws Bernie Sanders (senator from Vermont
who is not a member of any party)??
I don't like voting systems that make everybody be part of a party.
I think parties are evil. I wish they could be banned.
Unfortunately they cannot be, but *requiring* them is rather much.

>You know who your representative is, but they don't. All they know is that
>you're in one of their districts.

--this seems unhealthy. I mean, each rep has the attitude "golly,
everything I do is
flawless because every one of my constituents -- every single one --
always agrees with me.
Oh wait, some don't. OK... I deduce, those guys are just lying and
pretending to be my constituents.
OK... I conclude all that really matters are my money donors. I can be
sure who they are and that they really exist.
Screw the voters."


--I more or less understand your PAL system now (I think)
and it seems to me to be an extremely complicated kludge,
certainly one of the most complicated
PR system proposals ever advanced.

Lots of moving parts.
I'm not even sure it works; but it seems plausible it does (or can be
made to).
Also, the whole "nearby district" thing is kind of a mess.
Perhaps rulemakers will further step in to mess it up more to
manipulate things.

However, as you say, for the voter it may not seem so bad because
voter
just delivers approval-style ballot or a delegated-approvals bullet
vote.

So... I'd have to say I'm not very satisfied and certainly not
concluding "this field has pretty much been killed now."

I liked your initial insight/goal, which was that "single winner in
district" and
"proportional representation" do not necessarily need to be
incompatible.
That's a good place to start. But I hope one can cook up a lot
cleaner
system than PAL, to demonstrate that.


Jameson Quinn

unread,
Oct 28, 2011, 2:25:47 PM10/28/11
to electio...@googlegroups.com


2011/10/28 Warren D. Smith (CRV cofounder, http://RangeVoting.org) <warre...@gmail.com>



> > >The basic idea is:
> > >   - Candidates pre-announce their rank-ordering of the parties (starting
> >    with their own party) and may optionally approve/disapprove within
> > each
> >   party candidates.
>
> > --I cannot parse the above sentence after the word "optionally."
>
> They say for example "I like Red (candidates A and C but not B and D); then
> Green (all of them); then Blue (only candidate X)".

--aha.  Quite complicated.

> A vote is either a delegated bullet vote or an undelegated approval vote.

--aha.

> > >   - Each district "drafts" one member of each elected party from the
> >   elected slate.
> > --which one?
> See the section just below. This is just the outline, not the full rules.

> > >   - Your representative is the member of the party you voted for who is
> >   representing your district.
> > --so voters vote for parties not people???
> You vote for a person, that person has a party.

--so your system screws Bernie Sanders (senator from Vermont
who is not a member of any party)??
I don't like voting systems that make everybody be part of a party.
I think parties are evil.  I wish they could be banned.
Unfortunately they cannot be, but *requiring* them is rather much.

He's part of the Bernie Sanders party. This party could easily get a quota, or even more than one. (assuming of course that he were in the House, not the Senate; a proportional Senate requires either a new constitution or a series of two successive amendments).
 

>You know who your representative is, but they don't. All they know is that
>you're in one of their districts.

--this seems unhealthy.   I mean, each rep has the attitude "golly,
everything I do is
flawless because every one of my constituents -- every single one --
always agrees with me.
Oh wait, some don't.  OK... I deduce, those guys are just lying and
pretending to be my constituents.
OK... I conclude all that really matters are my money donors. I can be
sure who they are and that they really exist.
Screw the voters."

Ignoring invisible things that you know exist can be perilous. Or are you interested in a tour of scenic Fukushima? There's nothing visibly dangerous about it.
 


--I more or less understand your PAL system now (I think)
and it seems to me to be an extremely complicated kludge,
certainly one of the most complicated
PR system proposals ever advanced.

Lots of moving parts.
I'm not even sure it works; but it seems plausible it does (or can be
made to).
Also, the whole "nearby district" thing is kind of a mess.
Perhaps rulemakers will further step in to mess it up more to
manipulate things.

However, as you say, for the voter it may not seem so bad because
voter
just delivers approval-style ballot or a delegated-approvals bullet
vote.

So...  I'd have to say I'm not very satisfied and certainly not
concluding "this field has pretty much been killed now."

I liked your initial insight/goal, which was that "single winner in
district" and
"proportional representation" do not necessarily need to be
incompatible.
That's a good place to start.    But I hope one can cook up a lot
cleaner
system than PAL, to demonstrate that.


Well, that's what Balinski already did with FMV. But FMV is politically a nonstarter because it leads to "minority winners". As well as being far more party-centric than PAL.

The idea of PAL is to get 
1. single-winner-in-district 
2. PR
3. No arbitrary-seeming vote multipliers or minority winners
... by using explicit vote transfers to avoid minority winners. The rest of the rules followed from there. I would not be surprised if someone else could simplify it further, while preserving those three essential ends. 

JQ

Jim Riley

unread,
Oct 29, 2011, 2:19:04 AM10/29/11
to electio...@googlegroups.com, EM


On Wednesday, October 26, 2011 6:45:10 AM UTC-5, Jameson Quinn wrote:

2011/10/26 Jim Riley <jim...@pipeline.com>
I think that is an unrealistic interpretation of current law in the United States.

Thanks a lot for your legal analysis; you clearly know a lot more about this than I. I won't address your (very interesting) case-by-case discussion, but I will respond to your point about the politics of this:

I did a little more looking.  The requirement for election by districts was first enacted in 1842 as part of the Apportionment Act

"That in every case where a State is entitled to more than one Representative, the number to which each State shall be entitled under this apportionment shall be elected by districts composed of contiguous territory equal in number to the number of Representatives to which said State may be entitled, no one district electing more than one Representative"

I hope I copied this correctly from an image.  It appears that 1846/7 were the first elections where all representatives were elected from districts.  In then 1st Congress, about half of the representatives were elected at large, though in some cases, the representatives were from districts, while the voters were at large.  Larger States tended to use districts, as did new States.  Connecticut, Georgia, New Hampshire, New Jersey appeared to use at large election up until the federal law was passed, and in 1840, so did Mississippi and Missouri.  CT and NJ switched in 1842/3; GA in 1844/5; and the other 3 in 1846/7.

Since the intent was to eliminate the practice of at large elections, I don't think it can be interpreted as requiring districts, but somehow allowing election based on some PR scheme.  And there have been 10s of thousands elected from districts since then.  Because the number of districts is so closely tied to the apportionment of representatives, the laws tend to be expressed in terms of how and when the districts are drawn, rather than how the districts are to be used.

The "old" law before 1967 read this way.  Old is in quotes, because it is actually still part of the US Code (more about this below).  It is subsection (c), because (b) and (a) are dealing more with the actual apportionment and this is about the transition following a new apportionment.

2 U.S.C. § 2a(c)


(c) Until a State is redistricted in the manner provided by the
law thereof after any apportionment, the Representatives to which
such State is entitled under such apportionment shall be elected in
the following manner: (1) If there is no change in the number of
Representatives, they shall be elected from the districts then
prescribed by the law of such State, and if any of them are elected
from the State at large they shall continue to be so elected; (2)
if there is an increase in the number of Representatives, such
additional Representative or Representatives shall be elected from
the State at large and the other Representatives from the districts
then prescribed by the law of such State; (3) if there is a
decrease in the number of Representatives but the number of
districts in such State is equal to such decreased number of
Representatives, they shall be elected from the districts then
prescribed by the law of such State; (4) if there is a decrease in
the number of Representatives but the number of districts in such
State is less than such number of Representatives, the number of
Representatives by which such number of districts is exceeded shall
be elected from the State at large and the other Representatives
from the districts then prescribed by the law of such State; or (5)
if there is a decrease in the number of Representatives and the
number of districts in such State exceeds such decreased number of
Representatives, they shall be elected from the State at large.

So this is simply a transitional procedure following a new apportionment.  It is not a law permitting proportional representation that the 1967 law squashed. The timing of the 1967 law suggests that it probably was tied into the one man one vote decisions of the early 1960s, and possibly the Civil Rights movement.  At the time of the OMOV decisions, it was perceived that this would mean a big shift to the Democrats because rural seats would be dismantled and placed into the cities.  But by that time, they ended up getting moved to the suburbs.  1964 was a Democratic landslide (295-140), but that was because of LBJ-Goldwater and not redistricting.

It was understood at the time of Wesberry v Sanders that at large elections did comply with the Constitution since the representatives were chosen by the voters of their respective States.  In his dissent, Justice Harlan ridiculed the notion that Congress intended for districts to be of equal population, since so few representatives were elected at large, or from districts of remotely equal population.  And at the time of the 1967 statute, 6 federal courts were considering imposing at large elections as remedies for OMOV violations.

This is the section added in 1967, but 2a(c) was not repealed - it appears that there was some conflict between the Senate and the House of Representatives, and that was the reason 2c was enacted in an totally unrelated bill.  It wasn't a sneaky anti-PR bill.

2 U.S.C. § 2c


In each State entitled in the Ninety-first Congress or in any
subsequent Congress thereafter to more than one Representative
under an apportionment made pursuant to the provisions of section
2a(a) of this title, there shall be established by law a number of
districts equal to the number of Representatives to which such
State is so entitled, and Representatives shall be elected only
from districts so established, no district to elect more than one
Representative (except that a State which is entitled to more than
one Representative and which has in all previous elections elected
its Representatives at Large may elect its Representatives at Large
to the Ninety-first Congress).

The last sentence is a one time exception for Hawaii and New Mexico (the 91st Congress was elected in 1968.  New Mexico did not use the exception, Hawaii did, and then redistricted before the 1970 election.

After the 2c was added, it was accepted that 2a(c) had been implicitly repealed, since they seem to contradict each other.

But in Branch v Smith (2003) the US Supreme Court seemed to possibly maybe suggest that 2a(c) was still in effect.  After the 2000 census, Mississippi's apportionment was reduced from 5 to 4.  The legislature was not able to redistrict.  Eventually, a state court took over, and the Mississippi Supreme Court made a ruling that the lower state court could under the Mississippi Constitution redistrict.  There were lawsuits in both federal and state courts.  The federal court ruled that having a state court perform redistricting was a change in the baseline election procedures; and therefore had to be precleared under Section 5 of the Voting Rights Act (state court decisions have to be precleared, it is no different than legislation, or an executive decree, they are considered actions of the State).  Federal court decisions don't have to be precleared.  It also enjoined the State from seeking preclearance because it didn't appear that the Justice Department would preclear in time for the election.

This doesn't really have anything to do with 2c and 2a(c), but is background, and explains why most of the decision is about the VRA and jurisdiction of the state and federal court. 

Those who had taken the case to the federal court made the additional argument that though the federal court had jurisdiction to stop the state court, that they shouldn't have imposed their own redistricting plan, but rather should have applied 2a(c) and ordered at large elections.  I just listened to the oral arguments at the Oyez project.  One of the justices asked whether at large elections in Mississippi would be racially discriminatory.  He said that (1) since they would be enforcing a law passed by Congress (2a(c)) that the VRA does not apply; and (2) that they could enforce the 15th Amendment by ordering alternative forms of election - he seemed to imply cumulative voting.

This is the ruling of the district court.  Beyond the VRA argument, it makes the case that the Mississippi court did not have authority to create a redistricting plan.

http://www.senate.mn/departments/scr/redist/redsum2000/MS_301-CV-855WS_02-26-02.pdf

The US Supreme Court split 3 ways on whether 2a(c) applied.

3 justices (Souter, Breyer, Stevens) said that it didn't, and that 2c implicitly replied it.

2 justices (O'Connor and Thomas) said it did, since 2a(c) can be construed in a manner that they didn't conflict.

4 justices (including Scalia and Ginsburg) said that 2a(c) was still valid, but only in instances where there was no time for any other solution.  So basically it doesn't really apply since it would almost always be possible for a federal court to fashion some sort of districting plan.

But let's imagine that you convinced a State that Congress left a loophole.  You would likely then have a legislature where the party that is advantaged by district elections, is also advantaged in legislative elections.  So you go to Maryland or Illinois and tell the legislature that if they adopted this new system, that more Republicans would be elected to Congress. 

Very good point. A state legislature which performs a partisan gerrymander, will not want to undermine all that hard work by adopting FMV or PAL...
 
That their good friends who they had just contributed the maximum amount to, might lose his seat, and all their efforts at gerrymandering were for naught.   And in particular, you tell them that it is the white Democrats from swingy districts that will lose their seats.  If they were honest, they would tell you it was not to their partisan advantage.  But they could simply say that it doesn't appear to be legal and thank you for your time.

 ...and thus they will tend to (or pretend to) interpret any gray area of the law against the new system.

This is a pretty solid argument that FMV will not pass until the 1967 law is overturned.
 
If the 1967 law were repealed (it is unlikely to be overturned - since Congress has the authority of dictating the manner of electing representatives), then 2a(c) would remain in place.  But that doesn't necessarily get you much closer, since it doesn't permit a State to arbitrarily switch to at large elections.  You could only have full at large elections in States that lost representatives.  And you might get into problems if a State deliberately avoided redistricting (the transitional provisions are "until a State redistricts" under the "laws thereof".  Could a state lose a seat, and then change its laws so that redistricting doesn't happen, and then switch to at large.  Texas has no laws about congressional redistricting, so its "laws thereof" simply consist of the Texas Constitution vesting legislative authority in the Texas Legislature.
 
It is not an argument that the law will be impossible to overturn. Undermining partisan gerrymanders on a state-by-state level will tend to balance out more on a national level. And counting all the states with under 3 reps plus all the states with reasonably non-partisan redistricting rules (AK, AZ, CO, CA, HI, ID, IA, MO, MT, NJ, PA, WA), you can reach a majority of the House; so gerrymandered representatives don't have veto power over an overturn.

Arizona has a lot of controversy whether it was a non-partisan plan since the husband of the "independent" chairwoman was quite active in the Democratic Party, and she also ordered shredding of documents related to the hiring of consultants.  The representatives elected from those States might either favor retention of district elections since their State did such a wonderful job of redistricting; or they  might believe that they won despite the "fair" system trying to do them.  Their election from a district proves that election from districts is the best way.
 
Once the law is overturned, the fight will move to the states. There, it will be a tough fight, but there are two ways to overcome the pro-gerrymander bias Jim described. First, voter initiatives (in states which allow them); and second, states where partisan control has switched since the last redistricting. 

The reason for 2c was not to impose the use of districts; but to require States to switch to new districts immediately after an apportionment.  Election from districts has been the law for 169 years, and has been the general practice for 222 years.  You need to convince Congress to either require proportional representation or at least permit it.  It won't come from interpreting 2c as a sneaky anti-PR amendment.

Alternatively, would be to use the initiative power to switch to PR for a legislature.  The legislature will be wary because they were chosen from districts.  It works for them. 

Jim Riley

unread,
Oct 29, 2011, 2:48:45 AM10/29/11
to electio...@googlegroups.com
On Wednesday, October 26, 2011 5:50:31 PM UTC-5, Warren D. Smith (CRV cofounder, http://RangeVoting.org) wrote:
 
>More seriously, a mixed member system would be totally unacceptable to
existing incumbents, as it would draw too many of them out of their
existing districts.

--also [relatedly] highly unconvincing.  I think many politicians
don't give a damn that they are "drawn out of" their districts.  The
USA has in fact had in the past, states with "at large" congressmen,
so we know for sure you are wrong.

When SF had at-large elections for supervisor, they were slated and would run as a team.  It would be difficult to challenge them with an alternative team, because you would necessarily get some candidates who were more interested more in presenting an alternative viewpoint during the campaign, as opposed as presenting an alternative viewpoint when elected.  Others might have been passed over by the slate because they weren't effective speakers, etc.  So it might be enough to pull down the team.  If you ran as an individual, you might also require a weak member of the slate.  Otherwise voters would vote for 5 of the 6 members of the slate, and for you.  But they might skip around on which of the 6 not to vote for, and you might simply end up a strong 7th.  If you won, you might be added to the slate for next time.  When a member of the slate was going to run for another office, or retire, they might do so early enough that a new slate member could be appointed.  He would then run in  the next election as "Incumbent".

When SF switched to district elections, some supervisors didn't run, perhaps because they weren't a good match for their district, or would be in the same district as another supervisor.  Others were defeated.  An appointed supervisor finished 4th.

It is the change that the representatives would be concerned about.  They favor the system that elected them, over the system that didn't elect them.

Bruce Gilson

unread,
Oct 29, 2011, 5:42:16 AM10/29/11
to electio...@googlegroups.com
On first reading, I very much like the plan. As a practical thing, I cannot see it adopted for Congressional districts (except in those States that pemit initiative measures) because each State's legislature would not permit its power to gerrymander to be revoked. (In my State, Maryland, a Governor in 1990, horrified that this rather strongly Democratic State had a 4-4 division of its Representatives, came up with a grossly gerrymandered map, which the overwhelmingly Democratic legislature duly rubberstamped, that killed the seats of two of these Republicans. One of them, Connie Morella, who was popular enough in a normally Democratic district because she was rather liberal, found her district altered to include a very small, but overwhelmingly Democratic, piece of the next county over, where she was not known and the voters simply voted for the Democrat out of habit. She still won in those areas which had belonged to her old district, but her plurality there was not enough to overpower the monolithic Democratic vote in this new part. The other, Bob Ehrlich, decided there was no use running for re-election and had a victory of sorts: he ran for Governor, against the lieutenant governor of the previous governor, who was term-limited, and beat her statewide, despite the fact that the former LG was a Kennedy! Unfortunately, the State is pretty strongly Democratic, and four years later, Ehrlich could not get re-elected, though he was still quite popular. Fast forward to 2011 and the current governor found a way to eliminate one of the two remainig Republican seats, creating a very weird diatrict that ties the western, rural, Republican counties with a big enough piece of urbanized Montgomery County that it will probably have a Democratic plurality, even though these areas have nothing in common and nobody can really represent the whole district. This new map, of course, won't be tested till next November, so the plan may not accomplish its goal, but most people are convinced it will be.) It may be that the PAL plan can be implemented at a lower level than the statewide division into House districts, but I'm 69 years old and I suspect I will not see PAL adopted anywhere in my lifetime.
 
The one thing I see against it is the fact that it is so complex. I printed out Jameson's document and it is 9 pages long. I cannot see a plan that requires such a long explanation being adopted. Even the most complex plans presently in use anywhere, like the German/New Zealand MMP plans, can be summarized in a single page or less.
 
Too bad. The goals of Jameson's PAL system are worthy. Any plan which would give a voter like me a chance to have a representative who truly represented me is one I'd prefer over the current system. But for now it remains an intellectual exercise, and nothing more. 

Jameson Quinn

unread,
Oct 29, 2011, 10:20:32 AM10/29/11
to electio...@googlegroups.com, EM
Thanks again for the legal/historical analysis. Clearly, I'm out of my league. In fact:

This is a pretty solid argument that FMV will not pass until the 1967 law is overturned.
 
If the 1967 law were repealed (it is unlikely to be overturned - since Congress has the authority of dictating the manner of electing representatives),

Apparently I was even confused about the difference between repealed, overturned, and supplanted (though I get it now). I meant to say, any of the three. 

Thanks,
JQ

Jameson Quinn

unread,
Oct 29, 2011, 10:25:13 AM10/29/11
to electio...@googlegroups.com

It is the change that the representatives would be concerned about.  They favor the system that elected them, over the system that didn't elect them.

True. But in a place where the legislature happens to be properly proportional, you can argue that PAL will give the same results. They will be suspicious of this, but that's not the same as the the committed opposition that they'd direct at a more radically-different PR system such as MMP or multimember districts.

JQ

Warren Smith

unread,
Oct 30, 2011, 11:27:49 AM10/30/11
to electio...@googlegroups.com
I don't like the claim Bernie Sanders could still be electable because
he'd be the one
member of the "Bernie Sanders party."

It seems to me PAL severely disadvantages such candidates,
or anyhow effectively treats them highly differently.

They may already face disadvantages, but not ones intentionally built
into the voting system itself. The latter really rubs me the wrong
way.

Warren D. Smith (CRV cofounder, http://RangeVoting.org)

unread,
Oct 30, 2011, 11:36:55 AM10/30/11
to The Center for Election Science
http://www.rangevoting.org/CongHistory.html

discusses US congress elections and districts. Jim Riley was not
aware of
the fact there were a lot more "at large" congress members for a lot
more years.

Jim Riley

unread,
Oct 30, 2011, 3:53:00 PM10/30/11
to electio...@googlegroups.com, EM
I would interpret "overturned" to mean something that a court would impose, and "repealed" as something that a legislature would do.

Single-member districts have been the law for 169 years, and were the usual practice before then.  The 1967 law was intended to be a strengthening or clarification of the existing law.   It wasn't an anti-PR law.  At large elections are much less likely to produce a proportional result than single member districts.

The law is not going to be overturned by a court.  If courts were to start interpreting the old language in 2 USC 2a(c) as having any use, Congress would likely change the law; though they might ignore it, if it wasn't too much hassle to them personally (lose their seat).

I think it is pretty unlikely that FMV would be considered to be an election by district.  So you would have to convince Congress to change the law to permit use of FMV.   But it is too easy to calculate what would happen to individual members, for that to happen.

So you are still better off trying to introduce some sort of proportional representation at the state legislature level.  Since 49 legislatures have two houses, you might be able to switch a single house, and argue that it is equivalent to the election of the federal senate and house of representatives having different electorates.

Jameson Quinn

unread,
Oct 30, 2011, 6:41:42 PM10/30/11
to electio...@googlegroups.com
If an independent candidate for house of reps got 40% in their own district, 25% in the rest of their superdistrict, and a few percent in the rest of the state, they could win with no support from any partisan candidate. If they were the second partisan choice of some party, that would mean an average of half a quota extra - so those numbers could be halved.

I think that is fair, not discriminatory. PAL is actually much more party-independent than SPAL in practice, even though SPAL never mentions the word "party" and PAL does.

Jameson

2011/10/30 Warren Smith <warre...@gmail.com>

Bruce Gilson

unread,
Oct 30, 2011, 9:22:34 PM10/30/11
to electio...@googlegroups.com


On Sun, Oct 30, 2011 at 6:41 PM, Jameson Quinn <jameso...@gmail.com> wrote:
[...]

I think that is fair, not discriminatory. PAL is actually much more party-independent than SPAL in practice, even though SPAL never mentions the word "party" and PAL does.
 
 
 
What is SPAL? I haven't seen that one before. And I thought PAL meant "Proportion​al, Accountabl​e, Local," which does not mention the word "party." 

Jim Riley

unread,
Oct 31, 2011, 12:20:40 AM10/31/11
to electio...@googlegroups.com

Being that you are the mathematician, could you quantify "a lot" in terms of members and years?

For extra credit, you may express these values in terms of my awareness.

Regarding your History.

There were some States that always used at-large elections (NH, CT, NJ)  GA and RI fluctuated some.   PA and NY, and to a limited extent MD and MA  used multi-member districts in the larger cities (rather than attempt to split cities into districts).  For example, in 1840, NY elected 40 representatives from 33 districts, 28 were single member, and one elected 4; and 4 elected 2.  In PA 28 representatives were elected from 25 districts, 23 were single member, 1 had 3 members, and 1 had 2 members.  So it was essentially from districts, with a few areas not refined into districts.

Western states tended to use districts as soon as they had more than one representative.  Likely, legislators would want to have representatives for their own area, and i may have been a way to express pride in the growth of population.  There appears to be a few more States switching to at large elections in the 1830s, including AL, MS, and MO.  This might have been due to a heightening of sectional feeling before the Civil War, and the 1842 law might have been a response.

5 states, GA, MO, MS, NH, and IA used at large elections in 1842.  None did in 1846.  California entered the Union electing its 2 representatives at large.

After the 1842 legislation, at large elections were mostly transitional following a new apportionment.  States would gain new representatives which would be elected at large for the X2 election, and then the State would be redistricted by the X4 election.  At large elections for all representatives following a loss in an apportionment were fairly rare.  If a state gains a representative or two, and he is elected at large, the incumbent representatives don't mind.  Redistricting 25 districts into 26 will major disruptions to some districts if any attempt is made at equipopulous districts, so it may take time to develop the political will and agreement before redistricting goes forward.  Switching from 9 districts to 8 at-large seats will be a major inconvenience.

In the 20th century some large states went decades electing a representative or two at large.  They might even gain another and elect him at large.  Ohio appears to have stopped only when they lost the extra representative.

Louisiana used runoffs from 1978 to 2006, and has reinstated them for 2012.  Georgia also uses runoffs.

District elections have been the norm since 1788, and the law since 1842.  It is wishful thinking that the 1967 law was a sneaky behind the scenes attempt to kill off proportional elections.

Warren Smith

unread,
Oct 31, 2011, 2:13:25 PM10/31/11
to electio...@googlegroups.com
> Being that you are the mathematician, could you quantify "a lot" in terms of
> members and years?

--well, "Up until 1970, every congress but the 30th, 44-47th, 51st,
and 52nd contained at least one member elected 'at large' (i.e.
statewide not in districts)." Note the 52nd congress
was in the year 1891.

Pretty much everything on that page by me was got from the single book
Michael J. Dubin: US Congressional Elections 1788-1997, McFarland & Co. 1998.
Which is fine, but it means I'm sort of vulnerable. It would be better
to have more sources.

What sources do you use? It might be very helpful+interesting if you
(Jim Riley) were to write a post about how you do your research and
what are your sources, since you often seem to find out a lot of stuff
the rest of us did not know. Which is probably because
the rest of us don't know how to do that kind of research.

> District elections have been the norm since 1788, and the law since 1842.
> It is wishful thinking that the 1967 law was a sneaky behind the scenes
> attempt to kill off proportional elections.

--well, I'm not sure how to quantify "sneaky" & "behind the scenes"...
it was a rider attached to
a special immigration bill designed to help
Doctor Ricardo Vallejo Samala
immigrate.
Is that "in front of the scene" and "straightforward standard approach"?

I mean, doesn't that strike you as a teeny bit peculiar?


--
Warren D. Smith
http://RangeVoting.org  <-- add your endorsement (by clicking
"endorse" as 1st step)

Jim Riley

unread,
Nov 2, 2011, 3:34:00 AM11/2/11
to electio...@googlegroups.com

http://www.census.gov/history/www/reference/apportionment/apportionment_legislation_1790_-_1830.html



On Monday, October 31, 2011 1:13:25 PM UTC-5, Warren D. Smith (CRV cofounder, http://RangeVoting.org) wrote:
> Being that you are the mathematician, could you quantify "a lot" in terms of
> members and years?

--well, "Up until 1970, every congress but the 30th, 44-47th, 51st,
and 52nd contained at least one member elected 'at large' (i.e.
statewide not in districts)."  Note the 52nd congress
was in the year 1891.

I have already noted that the first requirement by Congress of single member districts was in the 1842 apportionment bill.

In the 27th Congress, elected in 1840/41 there were 31 representatives elected at large from 7 States, and another 11 extra elected from multimember districts in New York, Pennsylvania, and Maryland.  I counted a district with 3 members, as having 2 extra elected.   That was 42 of 243 not elected from single member districts.

In the 28th Congress, following the 1842 legislation, the multimember districts were eliminated, and there have not been any since.  3 of the 7 at large states had districted.  So 21 of 223 representatives were elected at large (the 1842 apportionment reduced the size of the House of Representative).  These States must have either ignored or defied the 1842 legislation, I don't know why.

In the 29th Congress, another State had come into compliance, so 13 of 223 representatives were elected at large.

And as you note, the 30th Congress was the first in which all representatives were elected from single member districts,  This was a direct result of the 1842 legislation.

Ever since that time, the number of representatives not elected from single member districts has been relatively small, and related to transitional provisions.

California became a State in 1850.  In 1849. California held an election for two representatives at large.  Historically, the People of an area have form a State government and then seek to accede to the Union (enter into the Constitution on an equal status with the other States), besides creating a state constitution, they elect representatives and senators.  Congress does not have the authority to grant statehood.   The territory of California had been acquired as a result of the Mexican Cession, and no census had been conducted.  The legislation accepting California into the Union doesn't specifically say that the representatives shall be at large, but that California should have two representatives.  It might not have been feasible to district.

There was not the usual apportionment bill following the census, but rather the apportionment method was specified in the census bill passed in 1850.  But there was a later special apportionment bill which said that California would have 2 representatives (this was because a fire had destroyed the census returns from San Francisco, and those from Santa Clara and Contra Costa were lost).  California continued to elect its two representative at large throughout the decade.

In addition, Minnesota joined the Union in May 1858.  At an election held in October 1857 in which 3 representatives were elected at large.  Congress only granted to seats, and one of the winners withdrew.  The population of Minnesota in 1850 was 6,000 and in 1860 was 172,000.  So there was no useful data for districting.

In September 1861, California elected 3 representatives to the 37th Congress, whose term began on March 3, 1861.  Until passage of the 20th Amendment, terms ran from March 3 of the odd year.  In addition, Congress did not typically meet until December, and meet through the following Spring, and sometimes into summer.  They would then have a short session from December of the odd year, until the end of their term in March.

Elections in the fall of even years, like we have now, really did not make sense.  The elected representatives did not take office for another 1/2 year, and often did not meet until over a year later.  And the outgoing Congress, including members who had been defeated would meet in a lame duck second session.  It might make some sense to have legislative elections in the fall.  Legislatures would meet in the winter (the most convenient time for farmers), but they would also have a start of term in the winter.

Many States would elect congressmen in the spring, just before the start of the term, or perhaps a bit later, so that representatives could run for re-election after the end of the lame duck session.  California waited a full 6 months, and as luck would have it, the Civil War broke out that Spring.  Lincoln called a special session in the summer, and California was not represented.  When its 3 representatives arrived, only two were seated.  But by March 1862, an apportionment had been made, and the 3rd representative was seated.  Minnesota also elected its 2 representatives for the 37th Congress at large.

The 38th Congress the first based on the 1860 census, California had 3 at-large representatives; Minnesota elected its 2 representatives from districts, and Illinois, which had gained a representative, elected him at large.  By the 39th Congress, California had switched to districts.  But Illinois continued to elect one representative at large.  I don't know why they did so, perhaps they argued that during the Civil War they didn't have time to bother.  During the latter years of the decade, some Southern states used at large elections for some representatives.  Many elections were conducted during this period and Congress did not always accept the results.

The 1870 census was the first after the elimination of slavery, which mooted the 3/5 rule for apportionment.  This meant that the Southern states would gain representation, even though on the losing side (the census missed a lot of people in the South, so the effect was somewhat muted.  Many Southern states gained representation following the 1880 census.  The 1872 apportionment law provided that States could use at large elections for added representatives.  A supplemental apportionment was made a few months later, to ensure no state lost representation.  It specified that the additional representatives could be elected at large for the next session (43rd only).  It is somewhat ambiguous whether this provision applied only to the supplemental apportionment or not.   The 1872 apportionment bill also established the uniform election date for Congress, required elections by ballots, required that new states have a population equal to the ratio of representation; and added the apportionment language of the 14th Amendment to statute.  It required that districts have equal population to the extent practicable, and explicitly outlawed multi-member districts (this had been true since 1842, but the statement in the 1872 law is more forceful).

In the 43rd Congress, 19 of 292 representatives, from 14 states were elected at large through the transitional provisions.

By the 44th Congress, all but Alabama with 2 at large representatives (and 6 at large had come into compliance).  Note that the summary in Dubin does not match the text.

For the 45th, 46th, and 47th Congresses, all representatives were elected from single member districts.

The 1882 apportionment bill included transitional language, and added a provision for states that lost districts, specifying that all representatives be elected at large.  8 states elected 12 additional representatives at large, and Maine which had lost a representative elected its 4 representatives at large.  So 16 of 325 were elected at large.

By the 1884 elections, all but Pennsylvania had redistricted.  So the 49th Congress saw one of 325 representatives elected at large (throughout this, I've ignored States with one representative, who can be seen as electing all members at large, or electing one member from a statewide district.   Pennsylvania didn't redistrict until the 1888 election to the 51st Congress.

South Dakota joined the Union in 1889 electing 2 representatives at large.  South Dakota had not been a separate territory in 1880. and had seen its population increase by 250% by the 1890 census.  So like California and Minnesota in earlier times, it didn't have the census data to really district.  Dubin is in error for the 51st and 52nd Sessions because of South Dakota.

ND, SD, WY, ID, MT, and WA were all admitted to the Union in the 1889 and 1890 (the 39th through 44th) and the first since Colorado in 1876.  With Idaho and Wyoming, they were contrary to the provision of the 1872 statute with respect to the minimum population for accession to the Union - but Congress can always pass new laws, and these States were likely to choose Republican representatives and presidential electors.

The February 1891 apportionment bill was somewhat earlier than in the past, and it may have have given more time for legislatures to redistrict.  In the 53rd Congress elected in 1892, there were 9 at large representatives from 5 states.  But unlike previous decades, there was little attrition, with 7 representatives from 4 states elected at large for the next 4 sessions (PA +2, KS +1, SD 2, and WA 2).

The January 1901 apportionment bill was also earlier than usual.  There 10 at large representatives from 6 states, but 6 representatives from 3 states were continuation of practice from the previous decade Kansas +1 and South Dakota did not have a changed apportionment, but continued to at large representatives.  Washington gained a 3rd district, but simply added the seat to the 2 already elected at large.

Colorado gained a 3rd representative in 1901, and elected at large.  In 2003, the Colorado Supreme Court determined that the state constitution mandated districting immediately upon gaining a representative and no other time.   After the 1910 census, Colorado gained a 4th seat, and elected that at large as well.  In 1913, Colorado created 4 districts, and modified them a bit in 1921 (they were malapportioned at the time they were created), and then didn't change them until the 1960s after Wesberry v Sanders.  I guess no one bothered reading the constitution for more than a century, or perhaps not at such a skewed angle.

4 States elected 6 at large representatives through the decade.

The August 1910 apportionment bill was somewhat later than the previous two, and in 1912 there were 27 at large representatives elected from 15 states.  Oddly, 3 states, CT, ND, and SD which had elected at large representatives for the entire previous decade, provided for all district elections in 1912, despite having no change in apportionment.

The number of at large representatives dropped quickly (10 at large from five states in 1914; 8 from 4 states in 1916; and 2 from Illinois in 1918.

There was no apportionment following the 1920 census, and Illinois continued to elect its 2 at large representatives for the next decade and beyond.

Congress finally apportioned based on the 1930 census (the apportionment was specified in the 1929 census bill.  Because of the 20 years from the previous apportionment, and the fixing of the number of representatives at 435, several states lost representation.    5 delegations with a total of 42 representatives were elected at large (13 in MO, 9 each in KY, MN, and VA, and 2 in ND).   All but North Dakota had redistricted by 1932.  There were another 12 representatives from 7 states elected at large.  Illinois did not gain or lose, but continued to elect the at large representatives they had since 1912.  The 54 at large representatives could be considered a lot.  The election of 1932 is not a lot, and is singular.

While the states that lost population and elected the entire delegation at large, quickly redistricted, But the others lingered on, and in 1948, 10 representatives from 6 states were elected at large.

The 1940s apportionment (made automatically - unless Congress had intervened).  Only added 3 states with at large representatives.  AZ and NM elected both of their representatives at large, and Florida elected the one that it had gained at large.  But the states that had at large representatives in 1940, continued to use at large representatives, with the exception of Oklahoma, which lost a representative, and therefore had the correct number of districts.  Ohio and Illinois lost a representative each, and went from electing two representatives at large, to electing one representative.  In total, in 1942. 12 representatives were elected at large in 8 states.

There was slow attrition through the decade.  In 1950 ND and NM each elect both (all) of their representatives at large, OH and CT had one at large representative, so there were 6 at large representatives from 4 states.  Illinois had 25 districts from 1902 through 1946, but from 1912 to 1940 elected two representatives at large, and from 1942 to 1946 elected one at large.  In 1948 it used 26 districts.

In 1952, ND and NM continued their practice of electing both representatives at large, and CT continue its practice of electing one representative at large (it used this scheme from 1932 to 1962).  Ohio's representation did not change, but they finally redistricted.  Texas and Washington gained a representative who was elected at large.  These two (TX and WA) redistricted by 1958, leaving 5 representatives elected at large from 3 states by the end of the decade.

The 1960 census cost Alabama its 9th seat, and in 1962 it elected 8 representatives at large.  New Mexico continued to elect its 2 at large, and Hawaii gained a 2nd representative which it elected at large (Hawaii should have had 2 representatives at statehood).  North Dakota finally created two districts.  MD, MI, OH and TX gained a seat each that was elected at large.  So that is 17 at large representatives from 8 states, with 8 of the 17 in Alabama.  By 1966, only New Mexico and Hawaii were the only at large states.  The 1967 permitted them to remain exceptional for one more election.  NM redistricted before the 1968 election, leaving Hawaii as the last state to have at large representatives.

So since the 1842 law, at large representatives have been exceptional.  There have been at large representatives in almost all Congresses for the next 136 years, but often just a couple and related to oddities such as new states acceding to the Union, or a fire destroying census records.  There was an uptick following censuses, and in the 20th century the laws provided that at large representatives could be elected "until the legislature redistricts", which was interpreted in relatively few states as "until Hell freezes over".
 

Pretty much everything on that page by me was got from the single book
Michael J. Dubin: US Congressional Elections 1788-1997, McFarland & Co. 1998.
Which is fine, but it means I'm sort of vulnerable. It would be better
to have more sources.

This is what I was using as well, which was why I was surprised that you said that I was unaware of the history.  The part you quote from is from the introduction, and doesn't quite match the results shown in the text.  Until I found that in the introduction, I thought you had simply been careless in compiling the data from the text.
 

What sources do you use?  It might be very helpful+interesting if you
(Jim Riley) were to write a post about how you do your research and
what are your sources, since you often seem to find out a lot of stuff
the rest of us did not know.   Which is probably because
the rest of us don't know how to do that kind of research.

Something that I had come across today was

http://www.census.gov/history/www/reference/apportionment/apportionment_legislation_1790_-_1830.html
http://www.census.gov/history/www/reference/apportionment/apportionment_legislation_1840_-_1880.html
http://www.census.gov/history/www/reference/apportionment/apportionment_legislation_1890_-_present.html

This has all the US apportionment laws.  Since apportionment is so closely tied to districts, it included all(?) of the laws requiring districts.   I'm sure you are familiar with Fair Representation by Balinski and Young.  If you look at most of the legislation it simply gives the end result (Maine gets 5 representatives, etc.)  So Balinski and Young had to research through congressional records and other sources to figure out how the apportionment was actually done.

A really really good source is A Century of Lawmaking for a New Nation

http://lcweb2.loc.gov/ammem/amlaw/lawhome.html

Somewhat limiting is that most of the text is in image form and not searchable.  But there are indexes from the original volumes that have been converted to searchable text.  I've just started to look at the history of the 1842 bill that first required districts.

> District elections have been the norm since 1788, and the law since 1842.
> It is wishful thinking that the 1967 law was a sneaky behind the scenes
> attempt to kill off proportional elections.

--well, I'm not sure how to quantify "sneaky" & "behind the scenes"...
it was a rider attached to
a special immigration bill designed to help
Doctor Ricardo Vallejo Samala
immigrate.
Is that "in front of the scene" and "straightforward standard approach"?

I mean, doesn't that strike you as a teeny bit peculiar?


I'll answer this in a separate post.

Jim Riley

unread,
Nov 2, 2011, 8:17:40 PM11/2/11
to electio...@googlegroups.com
On Monday, October 31, 2011 1:13:25 PM UTC-5, Warren D. Smith (CRV cofounder, http://RangeVoting.org) wrote:> District elections have been the norm since 1788, and the law since 1842.

> It is wishful thinking that the 1967 law was a sneaky behind the scenes
> attempt to kill off proportional elections.

--well, I'm not sure how to quantify "sneaky" & "behind the scenes"...
it was a rider attached to
a special immigration bill designed to help
Doctor Ricardo Vallejo Samala
immigrate.
Is that "in front of the scene" and "straightforward standard approach"?

I mean, doesn't that strike you as a teeny bit peculiar?

The US Supreme Court decision in Branch v Smith (2003) has quite a bit of the history.  (I usually use Findlaw for Supreme Court decisions  I think they have a complete set (since 1789)).
Justia and Oyez also have more modern decisions.  Oyez also has recordings of oral arguments, which were fairly interesting for this case.

There are really a lot of tangled issues in this case.  A couple of decades back, a black-majority congressional district was created in the Mississippi Delta, which is the area along the river south of Memphis, rather than the birdfoot delta south of New Orleans where the river flows into the Gulf of Mexico.  The first version was very ugly with lots of tentacles, because all blacks voted for the black candidate, and all whites voted for the white candidate, so you had to reach into areas with lots of Blacks and avoid areas with lots of Whites.  As time went on, the district boundaries became smoother, likely because some white voters have become more accepting of having a black congressman, and also because they don't want to have a head on white vs. black election.

Mississippi had 5 districts before 2000, when they lost one.  So they had to be careful how they increased the Black district from 20% of the state to 25% and also how they redrew the rest of the state, and which of the other two representatives would be paired.  If neither race nor incumbency nor partisan politics were involved, Mississippi would be relatively easy to redistictrict, beginning in the 4 corners of the state, with the SW area getting Jackson.

The legislature was unable to redistrict, and various parties went to court.  Or alternatively, the legislature did not redistrict because some legislators thought they would like the districts drawn by a court better.  Cases were filed in both Mississippi and federal court.  Branch v Smith is mostly about which court had jurisdiction.  The US Supreme Court has ruled that states, including the courts, have the authority to draw districts, and federal courts should only intervene as a last resort (see Growe v Emison).  Congress could draw congressional districts if it wanted to.  They have the authority under the Constitution to regulate the manner of congressional regulations.  They have used that authority to dictate use of single-member district elections.  But they have left the matter of where the districts are located up to the States.

After the legislature failed to act, some people sued in chancery court, which determined it had jurisdiction.  It appears that chancery courts are a fairly low level court.  They have a chancellor rather than a judge, and usually don't have a jury.

http://courts.ms.gov/aboutcourts/aboutcourts.html

The lead plaintiff, Beatrice Branch, is the first female president of the Mississippi NAACP.  So I suspect that the particular chancery court where the case was filed was one where they felt a plan favorable to blacks would be drawn.

Someone either didn't like their map or that they were drawing the map at all, and went to federal court.  The lead plaintiff was formerly the mayor of Meridian, and it appears that the plaintiffs were aligned with the Republican Party.  Mississippi is subject to Section 5 of the Voting Rights Act, which requires them to get clearance from the US Department of Justice for any change in election procedure, before it is put into effect.  The USDOJ never approves anything, but has 60 days to interpose an objection.  It is the obligation of those filing for preclearance to prove that the change does not have discriminatory intent or effect.  The USDOJ can request (demand) additional information from the state or locality, requesting additional information.  When that information is received, the 60 days is restarted.

Even if a redistricting plan is imposed by a state court, it must be precleared, because it is considered to be the state government making the change.  If a redistricting plan is imposed by a federal court, it doesn't have to be precleared.  So Mississippi sought to preclear the chancery court plan.  But then it was decided that Mississippi had to preclear the fact that the chancery court drew the map at all (the baseline procedure was that the legislature drew the map; and now someone else was doing it).   There is nothing in Mississippi law that gives explicit authority to redistrict to the chancery court in case of failure by the legislature.  So it appears that the chancery court was acting under a general authority to insure that the US Constitution be followed (in this case they were acting to prevent Mississippi from attempting to elect five representatives from non-equipopulous districts.  The Mississippi Supreme Court later ruled that the chancery court did have authority under the Mississippi Constitution to create a plan, but they never issued a formal opinion.  Meanwhile the USDOJ asked for additional information about the change in procedure to have the chancery court draw the map, which reset the 60-day clock.

The plaintiffs in federal court argued that the state court did not have authority under the US Constitution to redistrict.   Article 1, Section 4, "The Times, Places, and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof; ..."   The authority to prescribe is not vested in the State, but rather in the legislature.  There have been Supreme Court rulings to mean "legislative process", including the use of the referendum to block legislation, and the government veto.  A legislature doesn't have to actually draw the maps, but they would have to prescribe the process by which maps are drawn (eg they could pass a law that says a state court should draw the map in case that the legislature fails to do so).  It is still an open question whether the initiative can be used to prescribe the manner of electing representatives.

The plaintiffs in federal district court asked that the the court draw a map or alternatively impose at large elections (under 2 USC 2a(c), and also that they enjoin the state court from drawing its map.  The federal court enjoined the chancery court plan from being implemented because (1) it had not been precleared; and (2) it was unconstitutional for a state court, as opposed to the state legislature to draw a map.  They decided to not create at large districts but drew their own 4-district plan (which is still in use, BTW),

The state court plaintiffs (Branch et al) appealed the federal court decision, as did the federal court plaintiffs (Smith et al) on grounds that the court should have imposed at-large elections.   So the Smith plaintiffs got the case into federal court, but they didn't like the map.  The state court plaintiffs who had intervened in the federal court case, argued that if the state court plan not be used, that the federal court plan should be.

The US Supreme Court ruled that the federal court acted properly, because the chancery court plan had never been precleared USDOJ.  I suspect that the State of Mississippi must have been reasonably happy with the district court plan and stopped seeking preclearance.  The US Supreme Court vacated the part of the district court decision that it was unconstitutional for a state court to prescribe congressional district boundaries, barring any specific grant of authority by the legislature.  The supreme court ruled that it was premature for the federal court to look at the constitutional issue, once it had determined that the chancery court plan had not been precleared (and therefore there was no legal plan). 

The district court had essentially ruled, "it is unconstitutional for the state court to draw a plan, and besides this plan has not been precleared".  The supreme court said they should have ruled, "there is no precleared plan, and thus no plan (period)."  Because the Supreme Court ruled that consideration of the constitutional issue was premature, they were very vague about what the constitutional issue was,  It was if they had made the issue more clear, it would be as if they were considering the issue.   If you read the Kennedy concurring opinion, it explains this issue a bit more.

The Supreme Court decision then took up the issue of whether at large districts should have been imposed by the federal court.

This is discussed in Part III of the main decision; as well as the Stevens dissent, and the O'Connor concurrence/dissent.  It might make more sense if you start with the latter two, before trying to understand Part III of the main decision.  The Stevens dissent has much of the history of 2 USC 2c, which you find peculiar.

Congress has provided transitional provisions following an apportionment since 1872.  Because some States never transitioned, there have a few (not a lot) of at large representatives from 1842 to 1968, when they were finally eliminated.  2 USC 2a was implemented in the apportionment act of 1941 which definitively implemented Huntington-Hill and made reapportionment an automatic procedure conducted by the Census Bureau.  That is in 2 USC 2a(a) and 2 USC 2a(b) and is of course, still the law.

In 2 USC 2a(c) Congress prescribed the procedure for when the number of representatives changed, until the legislature draws new districts.

(1) If the number remains the same, they would be elected by the existing districts; and any at-large representatives would continue to be elected at large;
(2) If the number increases, the additional representatives would be elected at large, and any current at large representatives would continue to be elected at large.
(3) If the number decreases, but there are enough districts (ie there were some at large representatives) then the at large representatives would be be eliminated and the remaining legislators elected from districts.
(4) If the number decreases, but there are enough districts, and the number of representatives apportioned is greater than the number of districts, the excess can continue to be elected from districts (ie the number of at large representatives is reduce).
(5) If the number decreases, and there are not enough districts, then the entire delegation is elected at large.

There is not an explicit requirement to elect from districts.  However, in 1941, 38 States elected all their representatives from districts.  5 states (CT. IL, NY, OH, and OK) elected 103 representatives from districts, and an additional 8 representatives at large.  North Dakota was the only state with a multi-member totally at-large delegation (2 representative).  6 states (AZ, DE, NV, NM, VT, and WY) had single representatives (elected at large or equivalently a single statewide district).  So while a state could have all at large elections, they couldn't get to that status easily, since in general it required them to lose population share, and then have the legislature do nothing.  There have been only 5 instances where a state has lost representatives, and then had at large elections because of legislative inaction.  All of them managed to district by the following election Year XXX4.   This is in contrast to the case where states have gained representatives and elected a representative or two at large for decades.

As a result of the 1940 census, 5 states (IN, IA, KS, MA, and NE) lost representation and were members of Case (5), but created new districts before the 1942 elections.

3 states (IL, OH, and OK) lost representation, but had 1 or 2 at large representative.   OK lost its sole at large representative Case (3).  IL and OH, lost one of their two at large representatives case (4).

3 states (CA, NC, and OR) gained representation, but redistricted before the next election, avoiding Case (2).

3 states (AZ, FL, and NM) gained representation and elected at large representatives Case(2).  AZ and NM gained a 2nd representative, and both were elected at large.  FL gained a 6th representative, who was elected at large along with 5 district representatives.

3 states (CT, NY, and ND) maintained their number of representatives and continued their at large elections (1 at large from CT, 2 at large from NY, and both at large from ND).   The other 31 states with no change in representation continued to elect all representatives from districts.

Pennsylvania does not appear to have complied with the 1941 law.  In 1940, Pennsylvania had 34 representatives elected from districts.  It lost one representative.  But rather than creating 33 districts or electing 33 representatives at large, it elected 32 representatives from districts and one representative at large.  I don't know whether they ignored the law, flouted the law, or had some weird interpretation that reverted to some previous 32-district map.  In 1944, Pennsylvania did elect from 33 districts.

All these changes resulted 13 at large representatives from 9 states (AZ 2, CT +1, FL +1, IL +1, NY +2, OH +1, NM 2, ND 2, and PA +1)

By the end of the decade in 1950, this had been reduced to 6 at large representatives from 4 states (CT +1, OH +1, NM 2, ND 2)

As a result of the 1950 census, 7 states gained representatives and 7 states lost representation.  The 7 states that lost representation redistricted before the 1952 census, rather than hold at large election for the entire delegation (in New York, voters would have been instructed to vote for not more than 43).  5 of the seven gainers also redistricted, rather than electing an at large representative.  Texas and Washington did elect an at large representative in 1952, 1954, and 1956 before redistricting.

The 3 states which had elected at large representatives in 1950, CT, NM, and ND continued to do so through the decade.

As a result of the 1960 census, 8 states gained representatives, and 15 lost representatives.

Of the 15 losers, all but Alabama redistricted before the 1962 election.  Alabama elected 8 representatives at large, before switching to district elections in 1964.  Of the 8 gainers, 5 took advantage of the 1941 law to elect at large representatives rather than redistricting.  MD, MI, OH, and TX elected one representative at large, while Hawaii elected both of its representatives at large.  Meanwhile, of the 3 long term at large states, North Dakota switched to districts in 1962, and Connecticut to all district elections in 1964.

By 1966, only Hawaii and New Mexico elected their representatives at large.  Both are rather hard to redistrict, because it would likely require splitting of Honolulu and Albuquerque, and including territory 100s of miles away.  So there wasn't a need to get sneaky, since the law (2 USC 2c) has specific transition provisions for Hawaii and New Mexico (maybe they won't notice that they have to redistrict before 1970, and we can kick them out of the House of Representative?   Very unlikely).

What happened was Wesberry v Sanders in 1964.  Suddenly courts all over the country were inundated with lawsuits claiming malapportionment.  At first the courts were reluctant to get involved, and simply told the states to redistrict.  But the legislatures typically did not have the political will to make the necessary changes.  The legislative districts were often worse than the congressional districts.  There was an effort to have a constitutional amendment that would allow states to use something other than population as the basis for one house of the legislature.  Redistricting plans could get tangled up in procedural matters.  Or the plans might go half way, with perhaps 30% deviation (which is much better than 300%).  So the courts were pushed to get more involved.  They really didn't want to draw districts.  So as noted in Branch v Smith, federal courts started threatening to order at large elections.  These would address the issue of non-equipopulous districts.  But the States didn't want at large elections, and the representatives who were elected from districts didn't want to face election at large.  So they started considering how the laws could be changed.  There were proposals to permit a modestly large amount of deviation in district populations, at least during a transitional period.  Representatives from rural areas might favor this if they were realists, or oppose it if they somehow hoped that the issue of malapportionment would go away.  Representatives from urban areas hoped for a quick change, so that more like them could be elected.  All of these bills would have repealed 2 USC 2a(c), the transitional provisions following censuses.

But with no consensus in site, they managed to pass a provision requiring district elections, and used another bill that just happened to be handy.

The federal courts almost immediately backed off.  And everyone believed that 2 USC 2a(c) had been repealed.  That was until the issue was raised in Branch v Smith.  If you read the opinions you will see that it was a 4-3-2 split.  Justices O'Connor and Thomas said that it was still valid, and the district court should have imposed at large elections for all 4 representatives, rather than drawing a redistricting plan.

Justices Steven, Souter, and Breyer said that 2 USC 2a(c) had been implicitly repealed, noting that it included its own transitional provisions for the only two states with at large provisions at the time.

Justices Scalia, Kennedy, and Ginsburg, and Chief Justice Rehnquist argued that 2 USC 2a(c) was still valid, but then basically ruled there were no circumstances it could practically be used.  Mississippi had not redistricted.  So the federal court had to step in and draw districts.  But why, if Congress had provided the manner in which elections should be held in such circumstances?  That was because 2 USC 2c said elections should be from district.  So the only time 2 USC 2a(c) would be when the courts took too long to intervene.  But that is very unlikely to happen.

So it was a 6-3 decision that 2 USC 2a(c) was still in effect, and a 7-2 decision that 2 USC 2a(c) could never be used in practice.

Incidentally, the State of Mississippi has asked the district court (which has retained jurisdiction) to take over redistricting again this census, claiming that the legislature is unlikely to act.

And Mississippi which is one of fours states that has odd year elections for the legislature is the only one to fail to redistrict for 2011.  There is another court ruling, that they don't have to before 2012, even though they attempted to redistrict.  Since Mississippi elects its entirely legislature for 4 year terms, it is possible that they will continue to use districts based on the 2000 census until 2015.
Reply all
Reply to author
Forward
0 new messages