The redistricting process performed by Florida's heavily Republican legislature has been deemed unconstitutional. Considering the flagrant partisanship displayed by the legislature, and the state constitutional amendments which banned such behavior, it shouldn't be surprising, but the tortuous process of law means many people were breathing sighs of relief. I doubt they're doing that now.
The judge in this case, Terry Lewis, “ordered lawmakers to revise their congressional redistricting map” (Tampa Bay Times). The unacceptable maps drawn up by the legislature have now been revised. By the legislature. The results are predictable; very little change. And they might be accepted. Even if they're not, the Republicans will probably achieve their goal of deferring a just resolution until after the 2014 elections. They they'll work on eliminating state judges who disagree with their practices of election thievery.
Gerrymandering needs to die. That won't happen so long as the dance of the sugar daddies continues in the Republican state legislatures. I'll go further; I don't think it's possible to eliminate gerrymandering using the current approach. But there is another way.
Imagine that you walk into the voting booth this November. You are presented with a list of all candidates in a block of (usually) four districts1. Four districts, but you only have one vote. You denote your first choice – and second, third, etc. When you're finished, you hand in your ballot. That is one of only two changes which will be apparent to the voter. The other change is that party primaries are eliminated; they're not needed, and are no more than an opportunity for mischief. All the other changes take place in the processing of the votes – the software for which needs to be complete open.
It's called Single Transferable Vote, with multi-member districts. Using the existing districts, create district groups. Then, in the election, use the Single Transferable Vote method. (If you're unfamiliar with it, as I was, I urge you to go to the wiki link above).
These district groups are not your (profoundly unrepresentative) multi-member districts you may have read about in the past. These grouped districts with STV mitigate, or even eliminate, the effects of gerrymandered districts. They largely eliminate the leveraging of a sliver of registration advantage. And they provide an opportunity for minority groups to elect one of their own2.
I see numerous advantages to this approach.
- No immediate redrawing of district boundaries
- No guaranteed Republican (or Democratic) seats.
- Lesser advantage for incumbents, through the elimination of primaries. However, the larger geographical area could help the wealthier candidates.
- No votes thrown away when a candidate receives an oversized proportion of the votes.
- No votes thrown away when the votes are scattered among numerous candidates.
- No “plurality victories”. Quotes, because technically, all victories in this system are plurality victories (no one gets 50% of the total votes cast in the group of districts), but the situation of someone winning with fewer votes than the total of multiple opponents, does not apply.
- Minorities have an opportunity to elect a candidate of their choosing, without creating a single district composed of a majority of that minority.
- Fewer official lines separating urban centers and suburban/rural areas.
- More votes actually count. In a single district, up to 50% of all votes are for a loser.
In a 4-district group, up to 20% of all votes are for a loser.
Add to that the evaluation below, which indicates that multi-district STV would result in a fair result in Florida; 14D vs 13R, as close to the party registration ratios as you can get.
The ways votes are cast and counted make all the difference. We passed the wrong amendments in Florida. We should have chucked out the whole disgusting mess and replaced it with multi-member districts and STV. (and even numbers as much as possible, and completely open voting software).
(Everything that follows is boring technical detail elaborating on the use of STV)
If STV were used in Florida
The following maps display the Florida Congressional districts as of 2012, and those same districts in groups. Here is where the opportunity for partisan disrepresentation arises. The group maps consist of six groups of four, and one group of three. This yields, based purely on party registration, 14 Democrats and 13 Republicans.
Florida's districts, shaded for their Democratic or Republican advantage in registration, and the same districts in groups. (The green districts are virtually equal).
Current Districts with party bias |
Districts grouped, showing average party bias |
|
|
There are actually seven groups in the map to the right. I'm not proficient enough in Google maps to establish group borders, and the difference in registration is not enough to be apparent. See the footnotes for a table
3 with details.
And the districts showing the expected winner, based on which party has an advantage. For the grouped districts, winning parties are arbitrarily assigned to the districts in that group. So, for example, district 4 (top right) shows as a Democratic victory in the grouped district map, even though it is a majority Republican district, simply because it was the first in the list for that group.
Current Districts |
Districts grouped |
|
|
Here is the distribution of party representation as of 2012. It corresponds closely with that expected based on party bias. The lighter districts are those where party registration figures would suggest the opposite result; two, which happen to cancel each other out.
Current Districts, Actual 2012 |
|
|
Registration ratio: 53%D, 47%R
Representatives: 10D, 17R
Representatives ratio: 37%D, 63%R
vs, multi-member districts and STV
Registration ratio: 53%D, 47%R
Representatives: 14D, 13R
Representatives ratio: 52%D, 48%R
|
I look at those percentages in the box above, and think once again,
gerrymandering must die.
----------
How votes are cast (these next few items are all from the wiki article)
In STV, each voter ranks the list of candidates in order of preference. In the most common ballot design, they place a '1' beside their most preferred candidate, a '2' beside their second most preferred, and so on. The completed ballot paper therefore contains an ordinal list of candidates. In the ballot paper in the image on the right, the preferences of the voter are as follows:
1.John Citizen
2.Mary Hill
3.Jane Doe
How votes are counted
Setting the quota
In an STV election, a candidate requires a minimum number of votes – the quota (or threshold) – to be elected. A number of different quotas can be used; the most common is the Droop quota, given by the formula:
votes needed to win = (valid votes cast / seats to fill+1) + 1
where the quota is an integer. When the quota is not an integer it is rounded up; that is, its fractional part is increased to the next whole number. The Droop quota is an extension of requiring a 50% + 1 majority in single winner elections. For example, at most 3 people can have 25% + 1 in 3 winner elections, 9 can have 10% + 1 in 9 winner elections, and so on.
Finding the winners
An STV election proceeds according to the following steps:
1. A candidate who has reached or exceeded the quota is declared elected.
2. If a candidate has more votes than the quota, surplus votes are transferred to other candidates. Votes that would have gone to the winner go to the next preference. (So, if an immensely popular candidate receives 75% of the votes in a 4-member district (or, in the case of Florida, a district group composed of 4 districts), the votes above what's needed for victory (in this case, 20%) are assigned to the next choice. Different approaches can be used for this. I favor the Gregory method, using fractional votes).
3. If no-one new meets the quota, the candidate with the fewest votes is eliminated and those votes are transferred. (So, if the candidate you voted for gets very few votes, your vote is not just thrown away. It gets transferred to your 2nd choice. Or 3rd, etc.)
4. This process repeats until either a winner is found for every seat or there are as many seats as remaining candidates.
Footnotes
1. When I started writing this, I set up district groups in Florida with three 5-district groups, and three 4-district groups. I discovered that an odd number of districts in groups led to a disproportionate result, in this case 17 Democrats. Not only did the 4-district groupings eliminate that, but they act as a buffer, preventing a slight advantage from being inflated into a grossly disproportionate result for the state as a whole.
2. In a 4-district group, the percentage needed for victory is 20% (+1 vote). See Setting the quota, above. 20% should satisfy legal requirements to ensure minority representation, although IANAL. However, note that one of the groups for Florida is only 3 districts, since there are 27 Congressional seats. I don't see an alternative to this one 3-district group.
3. Table from spreadsheet with stats. ###### I'll set up a link to the info. I tried to plug it in here, and it resulted in a vast expanse of white space in front of the table. I'll add an update when it's ready.
5:08 PM PT: The table is in an addendum diary here.