Wisconsin Legislative Redistricting

[jimrtex]

Waukesha and New Berlin cities

[jimrtex]

Milwaukee city

[jimrtex]

I have completed the analysis for my districts for the 2004 presidential election, in which Kerry received 50.19% of the two-party vote.

Bush carried 57 districts, Kerry 42. The efficiency gap was 8.9%.

I was unable to attribute votes from wards that have no VTD associated with them:

Waukesha Kerry 36, Bush 36
Kenosha Kerry 80, Bush 88
Janesville Kerry 3, Bush 3
Madison Kerry 100, Bush 98
Oshkosh Kerry 0, Bush 5
Neenah Kerry 41, Bush 50
Green Bay Kerry 7 Bush 13
Eau Claire Kerry 185, Bush 76

Total Kerry 452, Bush 359

The missing votes in Eau Claire were from Ward 24. There are 38 VTD's for Eau Claire (1-39, with no 24). Either there is some reason that VTD 24 was not delineated, or they had skipped the number, and then backfilled after an annexation.

In the other cities the extra wards are at the end of the numbering sequence, indicating they were newly annexed areas which had been assigned new ward numbers.

Waukesha 17 new wards (9 with 0 votes, 63 of 72 votes in two new wards)
Kenosha 15 new wards (5 with 0 votes, 158 of 178 votes in three new wards)
Janesville 1 new ward (6 of 6 votes in that ward)
Oshkosh 5 new wards (2 with 0 votes, 5 of 5 votes in one new ward)
Neenah 8 new wards (81 of 91 in one group of three new wards)
Green Bay 1 new ward (20 of 20 votes in that one ward)
Eau Claire 1 missing VTD (261 of 261 votes in one ward corresponding to the missing VTD).

There are no missing votes from Milwaukee, New Berlin, and Racine, which appear to have more static boundaries.

Neenah reported its results by groups of wards associated with a polling place. This is possible because Neenah has fewer than 35,000 persons (around 25,000). I attributed the voting results to wards based on the VAP on a pro rata basis.

The efficiency gap is a measure of electoral parity.

Consider an area with a population equivalent to three districts that is 75% for the Center Party vs. 25% for the Moderate Party, we divide it into three districts that each also have 75% for the Center Party.

Wasted votes for Center: (75-50) * 3 = 75
Wasted votes for Moderate: 25 * 3 = 75

Net wasted votes 75-75 = 0 net wasted votes.

Now imagine that we pack the Center vote in one district and crack it in another such that the districts are 85, 75, and 65.

Wasted votes for Center (85-50) + (75-50) + ( 65-50) = 75
Wasted votes for Moderate 15 + 25 + 35 = 75

Net wasted votes 75-75 = 0 net wasted votes.

Let's make it even more extreme, with districts that are 95%, 75%, and 55%.

Wasted votes for Center (95-50) + (75-50) + (55-50) = 75
Wasted votes for Moderate 5 + 25 + 45 = 75

Now let's flip a district, so that support for Center is 95%, 85%, and 45%

Wasted votes for Center (95-50) + (85-50) + 45 = 125
Wasted votes for Moderate 5 + 15 + (55-50) = 25

Net wasted votes = 125 - 25 = 100 wasted votes.

If we make it 95%, 95%, and 35%

Wasted votes for Center (95-50) + (95-50) + 35 = 125
Wasted votes for Moderate 5 + 5 + (65-50) = 25

So only by flipping a district can we switch the efficiency gap measurement. And when we do so, we have a discontinuity, which is quite suspect in a measure of fairness.

For a complete legislature, we simply need to compute the number of districts with a Center majority (nC) and the average vote share in those districts (mC) and the number of districts with a Moderate majority (nM) and the average vote share for the Center party in those districts (mM)

Wasted votes for Center: nC * (mC - 50%) + nM * mM
Wasted votes for Moderate: nC * (100% - mC) + nM * (50% - mM)

Net wasted votes: 2 * (nC * mC + nM * mM)  - 50% (nC + nM) - nC * 100%

We can divide the expression by the total number of seats: nT = nC + nM, such that it represents the wasted votes per district, and substitute nC' = nC/nT and nM' = nM/nT (nC' and nM' are the relative share of seats).

Net wasted votes per district: 2 * (nC' * mC + nM' * mM)  - 50% (nC' + nM') - nC' * 100%

But nC' * mC + nM' * mM = mT, the average vote share for the Center party statewide, and
nC' + nM' = 1 (i.e. 100% seats are won by one or the other party).

Net wasted votes per district: 2 * mT - 1/2 - nC'

Solving for net wasted votes per district = 0.

nC' = 2 * mT - 1/2

Subtracting 1/2 from both sides of the equation.

nC' - 1/2 = 2 * (mT -1/2)

Substituting nC'' = nC' - 1/2  and mT'' = mT -1/2

nC'' = 2 * mT''

That is for the efficiency gap to be zero, then the excess share of seats above 50% should be twice the average vote share above 50%.

If a party gets 55% of the statewide vote they should have 60% of the seats. If a party gets 60% of the vote they should have 70% of the seats. If a party gets 80% of the vote they should have 110% of the seats (I doubt that happens even in North Korea).

All of the above are based on each district having the same number of votes cast. It would be rather bizarre if an election could be made fairer by varying the number of votes (or voters) in each district. This might be equivalent to giving weight to the percentage of land area covered in asphalt (which is just as bad giving voting weight to trees, see 'Reynolds v Sims').

[jimrtex]

What is the criteria for Solid, Lean, and Swing?

60%+
55% to 60%
45% to 55%
40% to 45%
40%-

??

[muon2]

What is the criteria for Solid, Lean, and Swing?

60%+
55% to 60%
45% to 55%
40% to 45%
40%-

??

Based on statistics from 2000-2010 Congressional races I use the following divisions:

PVI 0 or 1 tossup. Statistically the parties are equally likely to win in an even national environment.
PVI 2 to 5 competitive. Statistically the majority party will win about 75% of the time.
PVI 6+ uncompetitive. Statistically the majority party will win over 90% of the time.

Translating this to electoral margins in the two party vote one needs to adjust for the national average, since it can add or subtract a few percent. Then using your labels I would say

Swing: up to 51.5%
Lean: 51.5% to 55.5%
Solid: Over 55.5% (note that this is an 11 point margin or better)

[jimrtex]

This is the final statewide map. I have renumbered districts in a consistent fashion similar to that used in Ohio. The districts are numbered based on the population of the largest county within the district, so that 1-17 are in Milwaukee, 18-25 in Dane, 26-32 in Waukesha, 33-36 in Brown, and so on.



Western Wisconsin, details of St.Croix, Eau Claire, Wausau, and La Crosse



Eastern Wisconsin, details of Brown, Winnebago, Outagamie, and Sheboygan



Dane County, details of Madison



Mil-WOW-kee



Southeastern Wisconsin, Rock (Janesville), Walworth, Racine, Kenosha

[jimrtex]

1. Milwaukee: Milwaukee city (part, Wards 145-146, 184-189, 200, 216-234, 256-257)

Population: 0.990; 77% WVAP, 15% HVAP; Kerry (2004) 53.2%

2. Milwaukee: Milwaukee city (part, Wards 64, 135, 138, 142-144, 182-183, 235-236, 241-
249, 289-296) and West Milwaukee village.

Population: 0.993; 45% HVAP, 44% WVAP, 6% BVAP; Kerry 64.9%

3. Milwaukee: Milwaukee city (part, Wards 63, 132-134, 136-137, 139-141, 201-215)

Population 0.995; 68% HVAP, 20% WVAP, 8% BVAP; Kerry 74.9%

4. Milwaukee: Milwaukee city (part, Wards 16, 41, 48, 60-62, 65, 95-112, 297-298, 311-314)

Population 1.012; 48% BVAP, 41% WVAP, 5% HVAP; Kerry 83.7%

5. Milwaukee: Milwaukee city (part, Wards 66-73, 126-131, 181, 275-282, 299-310)

Population 1.011; 60% BVAP, 26% WVAP, 6% AVAP, 6% HVAP; Kerry 84.1%

6. Milwaukee: Milwaukee city (part, Wards 4-6, 11-19, 114-117, 122, 164-180)

Population 0.992; 92% BVAP; Kerry 95.1%.

7. Milwaukee: Milwaukee city (part, Wards 7, 9-10, 27, 29-37, 81-82, 84-87, 89-94, 113, 118-121, 123-125).

Population 0.996; 49% BVAP, 42% WVAP; Kerry 69.7%

8. Milwaukee: Brown Deer village; Milwaukee city (part, Wards 8, 20-24, 147-154, 156-163, 258).

Population 0.995; 57% BVAP, 33% WVAP, 5% AVAP; Kerry 72.9%

9. Milwaukee: Milwaukee city (part, Wards 25-26, 28, 74-80, 83, 155, 259-261, 263-273).

Population 0.995; 53% BVAP, 36% WVAP, 5% AVAP; Kerry 70.2%.

10. Milwaukee: Oak Creek city; South Milwaukee city.

Population 0.968; 87% WVAP, 5% HVAP; Kerry 47.5%

11. Milwaukee: Franklin city; Greendale village; Hales Corners village.

Population 0.996; 88% WVAP; Kerry 39.8%

12. Milwaukee: Cudahy city; Milwaukee city (part, Wards 54-59, 237-240, 250-255); St. Francis city

Population 1.008; 86% WVAP, 6% HVAP; Kerry 59.7%

[jimrtex]

This is a study cited by the plaintiffs in their brief to the SCOTUS

The Impact of Political Geography on Wisconsin Redistricting: An Analysis of Wisconsin’s Act 43 Assembly Districting Plan (PDF)

The Chen study cited by the brief illustrates that the "Efficiency Gap" does not measure what it purports to measure. If we look at Figure 2, we will notice that the results line up in columns - it is not a continuous function.

If we compare Figure 2 with Figure 5, we will see that the two share the identical pattern (the X axis is flipped, in Figure 2 more Republican results are to the left, while in Figure 5, the more Republican results are to the right.

The efficiency index is in reality a measure of proportionality. The fuzziness of the columns in Figure 2 is solely due to the variation in the number of votes cast in a district. According to the proponents of the Efficiency Gap, half the votes in a district are "wasted". If there are more votes cast, there are more wasted votes.

Thus according to advocates of the Efficiency Gap, a plan could be made fairer or unfairer by placing more children, non-citizens, felons, low-participation voters in some districts - which hardly seems consistent with the ideal of one voter, one vote.

[jimrtex]

The Supreme Court has agreed to hear the case on appeal, which will be styled Gill v Whitford. The court also stayed the redistricting order on a 5-4 vote, with the four Democrat judges voting to let the redistricting to go forward.