Tuesday, August 13, 2013

Was Fairbanks Gerrymandered? Riley Challenge to Alaska Redistricting Board's 2013 Plan Part 1

[I started on this a couple of weeks ago, when I got a copy of the Riley challenge to the Redistricting Board's newest plan.  Since then the Board has also issued a response which you can see here.  Basically it affirms that the earth revolves around the sun, but denies most everything else. I'm not going to discuss the Board's response in this post.  I got it late and the post is already too long.  My point is to help people understand the challenges, but not decide the case.]

In this post I'm going to list some of the key allegation made by the plaintiffs about the redistricting plan along with their complete filing.  In this post I'm focusing just on the Fairbanks part of the challenges. In the end I left some things out and some in though they are speculative. 
  1. The specific problems with the Fairbanks districts listed in the Riley court challenge.
  2. Explanations of the terms (in red) 
  3. Fairbanks maps so you can see the districts
  4. Discussion of the claims made by Riley, though I'm leaving the truncation issue to a later post. 
  5. The complete court document

1.    Specific Problems  (I'm quoting here from the Riley challenge; I've added the red.)

14.  The Third Final Plan establishes the boundaries of House Districts [sic] 3 which fail to comply with Article VI, Section 8 of the Alaska Constitution in that it is not relatively compact.


15.  The Third Final Plan establishes House Districts 4 and 5, which unnecessarily divides the campus of the University of Alaska and fails to comply with Article VI, Section 8 of the Alaska Constitution by drawing a boundary that unnecessarily divides an area that comprises an integrated socio-economic area. 

16.  The Third Final Plan fails to comply with Article VI, Section 8 of the Alaska Constitution by establishing House Districts 1-5 with unnecessarily higher deviations from the ideal district population and do not contain populations as near as practicable to the quotient obtained by dividing the population of the state by forty. 


17.  The Third Final Plan establishes Senate District B which is  unnecessarily non-contiguous and non-compact territory as required by prior Alaska Supreme Court cases.

18.  The Third Final Plan fails to comply with Article VI, Section 8 of the Alaska Constitution by drawing boundaries with unnecessarily higher deviations from the ideal district population and violates equal protection of voters rights to an equally weighted vote and the right to fair and effective representation.


25.  The Board's Truncation Plan for Senate Districts improperly considered improper factors (a) substantial changes from an unconstitutional Interim Plan as opposed to the prior Final Plan in effect for the 2010, b) incumbency protection relative to Senate District B; and (c) previously considered partisan voting patterns of persons within the Ester/Goldstream Area. 
2.  Explanations of Terms

What does this mean?  The US and Alaska constitutions require one-person-one-vote, thus there should be minimal deviation between districts.  The Alaska Constitution requires that house districts be compact, contiguous (all connected, not separated), and socio-economically integrated.

Deviation - The one-person-one-vote principle means that every district should be equal in population.  The 2010 Alaska Census counted 710,200 Alaskans.  There are 40 House districts.  Dividing the population by 40 yields 17,755 people per district as the ideal.  Deviation is the number (or percentage) of people in a district above or below that ideal number.  In the urban areas the goal is to keep deviations at or below 1% (177 people).  In the rural areas, in the first rounds of this process, when the Voting Rights Act required pre-clearance from the Department of Justice, the Board followed a rule that 10% was an absolute maximum allowable (but to be avoided if possible) total deviation in the state from the district with the highest positive deviation to one with the lowest negative deviation.

The general rule is that deviations should be as low as possible to achieve the other goals.

Compact - Districts should be as small and concentrated in area as possible.   One problem in achieving this in Alaska is that we have a lot of land and relatively few people.  Since all the districts need the same number of people, where there are lots of people (urban areas) it's easier to draw compact districts.  In rural areas it is harder to find 17,755 people and also keep the district compact.  It could get larger and/or have strange protrusions to capture villages here and there to get the population up.

Here's a table to show what this means more visually.

Examples of District Maps House Districts Senate Districts

6 imaginary districts, each compact and
Made up of two house districts.
Ideal options =
1&2      1&2     1&3
3&4      4&6     5&6
5&6      3&5     2&4  
(The Board wanted the Senate districts to be made up of House districts in numerical order - 1&2, 3&4, etc.)

Technically, 1&4 could
be contiguous, but would be questionable and then 2 & 6 would be isolated. 

6 is the only  compact district. 3 & 5 have those extensions. 1, 4, & 2 are ridiculous.

1 and 4 are not contiguous. 3 is iffy.
The pairings would have to be as compact and contiguous as possible.  But I think the House districts are so bad, it would be impossible to create constitutional Senate districts.

Contiguous - Basically the districts should be one area without any breaks. In the squiggly map in the table, districts 4 and 1 are not contiguous, because there are areas not connected to the rest of the district.

In Alaska, there are islands connected to other land over water. They called this contiguous over water.  But the islands were too small to be a whole district.  See House District 32 in the new plan.  HD 32 connects Kodiak to Cordova and Yakutat.   When the Board decided to use the AFFER Matsu map instead of the Calista map, they were troubled that the Calista map connected north Anchorage to the rest of the district (Valdez and the pipeline corridor) over glaciers and uninhabited mountains. 

These two criteria are relatively easy to determine.  If you start with a squarish district, you may need to stretch it or have protrusions off the square to get a pocket of population here and there.  The question down the line will be whether those deviations from a tight compact district were necessary to meet other criteria, reflect geographic features (like a meandering river), reflect quirks in the census blocks,  or are done to include or exclude particular people or groups of people. 

Socio-economic integration - This is a little harder and more abstract.  In Alaska, keeping political units together is important.  A city, like Anchorage, or borough like the North Star Fairbanks Borough, is considered socio-economically integrated.  So combining Muldoon and Eagle River is considered ok since both neighborhoods are in Anchorage, even if they are economically different.  (They were combined in the previous plan, but not this one.)

The Board's job was to create equal  districts (minimal deviation) while balancing compactness, contiguity, and socio-economic integration.  When judging that, the Court has to determine if anomalies were due to the geography and population distribution or attempts to gerrymander.

Truncation:   Senate terms are for four years, while house terms are for only two.  Senate seats are also staggered.  Half (10) are voted on in one election and the other half (10) in the next election two years later.  If redistricting significantly changes the constituency of a senate seat, then a large number of the voters of the new district are represented by someone they didn't vote for.  Thus, senate seats with significant changes are subject to truncation.  This means that regardless of when the term is up for the sitting senator, the population should be able to participate in choosing their senator in the next election.

So, all the new districts whose terms expire in 2016 with a significant change will be up for election in the next election (2014).  Those up for election in 2014 will be up again anyway so they don't need to truncate.  But this messes up the staggered terms, so some have to be designated as two year terms and others as four year terms to get ten up for election one year and the other ten the next election. The 2012 election used a new redistricting plan in which all but one of the seats were truncated and then the Board assigned two or four year terms to them. And now they have to do that again. 

3.  Looking at Fairbanks

Click for bigger and clearer map
The numbers indicate the House District and the letters indicate the Senate District.  Two contiguous house districts make up one senate seat.  

The map shows districts 1, 2, and 3 completely and parts of districts 4, 5, and 6. In fact, HD 5 is large, and HD 6 is huge.  Here's a map that shows all of 4, 5, and 6.

Unfortunately the colors switch from map to map.  Fairbanks is in the center.  You can see 3B, 4B, and 5C.  (1A and 2A are too small to see in this map.)  House District 6 is that huge sea of blue along the Canadian border, around Fairbanks and back down the other side.  You might also note 9E.  It goes from Fishhook Road near Wasilla, to Valdez and then up the pipeline corridor to the edges of Fairbanks. 

4.  The Riley Challenges regarding Fairbanks

The Riley challenge's first complaint (14 in the court document) is that "House Districts [sic] 3 is not compact."  While it doesn't look all that big, HD 3 is long and stretched out, so they are claiming that people are further from each other than is necessary.  If this were the only available population in the area, this might be unavoidable, but I'm told there were plenty of people available to make a more compact district.

Yellow is HD 5 and dark blue is HD 4
The second complaint (15)  is that the University of Alaska Fairbanks (UAF) is unnecessarily split between two house districts (4 and 5).  Political units are supposed to be kept intact if possible, but I'm not sure that the university campus qualifies.  Yet, common sense would keep the campus together unless there was a compelling reason to split it.  College campuses tend to vote more liberally than the general population.  I was told that the two precincts  (in the new 5) that voted Democratic in the 2012 election have been disrupted. One was put completely in the already Democratic leaning District 4 and the other (the University) was split between 4 and 5 . I don't know Fairbanks and I had trouble matching up the Board's maps to University maps, but I think the map gives a reasonably close approximation of where the University is.

I don't know where the dorms are and how they are split up, if at all.  And I don't know how many students register with their University address rather than their home address.  So I don't know how many actual voters are affected.

The third issue (16) raised is that districts 1-5 have unnecessarily high deviations.   Let's look at the Fairbanks deviations.  I've included HD 6.

House District Senate District Total Population Percent Deviation
From Ideal
# Deviation
17,726 -0.16% -29
17,738 -0.10% -17

A 35,464 -0.13% -46
17,673 -0.46% -82
17,786 0.17% +31

B 35,459 -0.14% -51
17,837 0.46% +82
17,807 0.29% +52

C 35,644 0.38% 134

The individual districts are all well under one percent deviation.  The total deviation from the lowest (-0.46%) to the highest (+0.46%) does come to 0.92%.  But that is still under one percent.  Anchorage, the biggest urban area (it's easier to have lower deviations where there are more people,)  has higher deviations.  On the face of it, I think these deviations should be fine.  UNLESS, it's clear that they could easily have been made lower, combined with other issues like compactness or contiguity or gerrymandering.    But gerrymandering hasn't been raised. (Well, not exactly true.  The word wasn't used but the truncation challenges suggest political intent.) And showing intent is pretty hard. 

The next two issues apply to Senate districts.

The fourth (17) is that Senate district B is unnecessarily non-contiguous.  HD 4 is a pretty large district (the first map above only shows part of it) and it's only connected to HD 3 at one little 2.5 mile spot that doesn't appear to have much population.

2.5 mile connection between HD 3 and 4

In contrast, much of HD 4 is connected in a long swath to HD 5 and they have the University split between them.  As it is, HD 5 is a strangely drawn district.  Most of it - I'm told all of it below the river - is uninhabited military bombing range.  Nearly all of the populated area is west of the City of Fairbanks.  There's a tiny jigsaw piece to the east of the City.  It doesn't appear to have much population.  And it looks like it's only contiguous with the rest because of the bombing range.  But I don't think one could drive to the main part of the district without going outside the district.

I understand that the courts have said contiguity doesn't require that people are able to drive from one part of the district to another.  But I suspect that ruling refers to rural districts with villages not connected to the road systems where it's hard to find enough population for a district.  Downtown Fairbanks is an entirely different situation.  

Here's the map of District 5. In the larger scale, the map's cut off on the left.  The inset has the whole map but it's tiny. (If they can't make a map with the whole district, does it mean it isn't compact?)

The Fairbanks News Miner has editorialized that the board should have paired HD 4 and HD 5 into one senate district  and HD 3 and HD 6 into another.

As I look at this, it seems like a reasonable idea.
  • There's a long border between HD  5 and  HD 4 with connected neighborhoods.
  • Pairing 4 and 5 would  reunite the university in a single senate district.  (Light blue circle.)
  • Most of HD 5 is uninhabited bombing range and essentially the eastern part of HD 5 (big red circle on the lower right) is NOT contiguous with the west part in any real sense.  I don't think you can drive from one side to the other without going out of the district.  (Maybe you can go by boat along the river.)
  • The real border between the populated area of HD 5 and HD 6 is a tiny little corridor. See the circle in red in the inset with an arrow pointing to where it would be if they showed the whole district on the map.  

And if you look at the deviation table, you'll see that HD 5 has 82 too many people and HD 3 has 82 too few people.  I wondered how many people lived in the east pocket of HD 5 and whether just giving that pocket to HD 3 would balance them.  Well, I was told there are about 500.  Too many. But having watched the Board move around population on the computer to find better borders, I'm convinced that there's a way to make some adjustments to get rid of this de facto non-contiguous pocket of voters.

But as it stands Senate district B (3&4) has a deviation of -51.
Senate district C (5&6) has a deviation of 134.

If you paired 3&6, the new senate district deviation would be +30.
The deviation for new senate district of 4&5 would be +113

Combined, the deviations would be lowered by 42 people.  That by itself is not much, but combined with all the other issues, it seems like these two senate seats were mispaired.

It also appears that  a senate pairing of HD4 & HD5 would have a greater chance of electing a Democratic senator than the way the Board paired them, which would be a good reason for some on the Board to prefer the HD3 & HD 4 and HD 5 & HD 6 pairings. 


There was one more Fairbanks related issue, truncation.  I think this post is already long and confusing enough without adding the truncation piece.  While the topics here are all very closely related, truncation is really a different issue and can be handled separately.  I'll do that in another post. You'll see that the Board was pretty spacy by that time.  I did post on the truncation Board meeting already for those who can't sleep without knowing more about this charge.  And that post links to a post two years earlier where I tried to explain truncation when it came up with the first plan. 

Below is a copy of the Riley challenge to the Board's most recent final plan.  As I mentioned at the beginning, the Board has replied to this challenge point by point and you can read that here.

Riley Challenge To July 2013 Alaska Redistricting Plan

PART II on Truncation is here.


Comments will be reviewed, not for content (except ads), but for style. Comments with personal insults, rambling tirades, and significant repetition will be deleted. Ads disguised as comments, unless closely related to the post and of value to readers (my call) will be deleted. Click here to learn to put links in your comment.