Apportionment fun

[Beefalow and the Consumer]

If anyone is interested, I wrote a Perl script to take a file containing states and populations, and figure out how many EVs each state gets, using the current method of Congressional apportionment.  Some interesting finds:

Utah was 856 residents away from stealing an EV from North Carolina in 2000.

According to the 2010 population projection made in 2001 by the US Census Bureau, the following EV changes will take place in 2010:

CA +2
TX +2
WA +1
OR +1
UT +1
MT +1

NY -2
IL -1
PA -1
OH -1
MI -1
MA -1
IA -1

I was surprised to see CA gaining an extra two EVs, while AZ, CO, GA, VA, and TN didn't gain any.  Florida came very close to getting another EV at Texas' expense.

Last five EVs awarded:
5. Missouri (11) (falling)
4. California (57) (rising)
3. Minnesota (10) (falling)
2. North Carolina (15) (static)
1. Texas (36) (rising)

First five EVs almost awarded:
1. Florida (28) (rising)
2. Michigan (17) (falling)
3. California (58) (rising)
4. Massachusetts (12) (falling)
5. Illinois (21) (falling)

So, Minnesota and Missouri are also in danger of losing EVs, but are not projected to lose them in 2010, based on the population projections I found.

GOP states are looking to pick up 4 EVs, with Democrat states losing 1, and Battleground states losing 3.

[Colin]

Ya go Montana. 4 EVs.

[patrick1]

If anyone is interested, I wrote a Perl script to take a file containing states and populations, and figure out how many EVs each state gets, using the current method of Congressional apportionment.  Some interesting finds:

Utah was 856 residents away from stealing an EV from North Carolina in 2000.

According to the 2010 population projection made in 2001 by the US Census Bureau, the following EV changes will take place in 2010:

CA +2
TX +2
WA +1
OR +1
UT +1
MT +1

NY -2
IL -1
PA -1
OH -1
MI -1
MA -1
IA -1

I was surprised to see CA gaining an extra two EVs, while AZ, CO, GA, VA, and TN didn't gain any.  Florida came very close to getting another EV at Texas' expense.

Last five EVs awarded:
5. Missouri (11) (falling)
4. California (57) (rising)
3. Minnesota (10) (falling)
2. North Carolina (15) (static)
1. Texas (36) (rising)

First five EVs almost awarded:
1. Florida (28) (rising)
2. Michigan (17) (falling)
3. California (58) (rising)
4. Massachusetts (12) (falling)
5. Illinois (21) (falling)

So, Minnesota and Missouri are also in danger of losing EVs, but are not projected to lose them in 2010, based on the population projections I found.

GOP states are looking to pick up 4 EVs, with Democrat states losing 1, and Battleground states losing 3.

I have done similar calculations.  Beef something is wrong with your formula.  Florida will definitely get one electoral seat and most probably will get two.  Texas will get 2 and 50/50  3.  Washington Oregon and Montana will not get seats.  Baseline for a house seat might be around 766 thousand in 2010.  Montana will not get another seat.  Georgia, Nevada and Arizona will gain seats based on census projections.  I essentially agree with your lose seats. Most projections show OH losing 2.  Minnesota is will probably lose 1 also.  Census projections mark Louisiana for a loss.  Alabama is in danger and WV might be if they dont start pumping out some kids.

[muon2]

We had a thread on this in October. My analysis from that thread is quoted here.

I took the 2000 census and the 2003 update from the census. If I project the growth between 2000 and 2003 to April 1, 2010 I get the following changes to Congressional apportionment.

AL -1 (8 EV)
AZ +1 (11 EV)
CA +2 (57 EV)
FL +2 (29 EV)
GA +1 (16 EV)
IL -1 (20 EV)
IA -1 (6 EV)
LA -1 (8 EV)
MA -1 (11 EV)
MO -1 (10 EV)
NV +1 (6 EV)
NY -2 (29 EV)
OH -2 (18 EV)
PA -1 (20 EV)
TX +3 (38 EV)
UT +1 (6 EV)

The last few seats I apportioned were:
432 - CA 55
433 - PA 18
434 - MN 8
435 - MI 15

The next in line were
NY 28
IL 19
AL 7

Then after a gap
CA 56
MD 9
OH 17
TX 36

It's pretty easy to move any of the last four out and the next three in. Before year's end, I expect the Census Bureau to release the July 1, 2004 estimates. I'll rework my projections at that point.

[True Federalist (진정한 연방 주의자)]

Montana will not get another seat.

Montana came close to getting another seat this time patrick.

The priority value for the 435th seat assigned in 2000 was 645,931 and that for Montana getting a second congressional seat was 640,155.  The priority values, which are used to detemine whether a extra seat gets added someplace is for the n^th seat as follows:
   population / sqrt(n*(n-1))
So to get a second seat a state needs apprixmately 70% of its apportionment population to be near the per district average, not 50% of its population.
  Based on the Census estimates of population, it looks like the priority value for a 2nd Montana CD in 2010 will be 700k +- 5k.  Average approtionment population per district will be approximately 690k.  I will be extremely surprised if Montana does not gain a CD as a result of the 2010 census.

[Bogart]

This is what I come up with using Census projections and the current method of apportionment. That being as follows:

1. Each state automatically gets one seat.
2. Calculate mulipliers for each "next seat" by dividing 1 by the square root of the next seat times that seat minus 1, or 1/sr of n(n-1).
3. Take multipliers by each states projected population to obtain priority values for gaining each additional seat.
4. Rank them in decending order.

It has been pointed out to me that some Census projections are weighted too heavily towards California growth and this may be true. However, based upon these projections, this is how I calculated it turning out for 2010.

[True Federalist (진정한 연방 주의자)]

Cleaned up the map slightly


And here's a version with DC and PR added:

[??????????]

Ernest,

Those maps are just gain/loss of congressional seats. You are not including senate seats. Correct?

[Alcon]

How does adding Puerto Rico and DC make Nebraska and Maine gain two seats each?

[khirkhib]

I wonder where they would put a new congressional district in Oregon.  I think with PR and DC added does total up right.  No state lost any Reps.  There could be quite a lot of new congressional districts to make in the 2012 elections.

[True Federalist (진정한 연방 주의자)]

How does adding Puerto Rico and DC make Nebraska and Maine gain two seats each?

Apparently the version of Dave's map generator that I am using that has PR available for display has the number of votes displayed for Maine and Nebraska hardcoded to their EV's.  That's how they gained two seats each.

[Beefalow and the Consumer]

Cleaned up the map slightly

Looking over this, it looks pretty much like my prediction, except CA should be 56 reps instead of 58.  (Ah, I think one comes from Ohio and the other comes from Minnesota.  And another Ohio seat is going to Florida - you must have very different numbers than I have.)

Anyone have some more recent state population projecitons for 2010?  I can feed those numbers into my program (I really ought to create a Web interface for this...)  I think it's going to be a close battle between TX and FL for the 435th seat.

[Dave from Michigan]

Michigan's going to lose a seat again.  We've been losing them since 1980 I think 

1980  -1
1990  -2
2000  -1
2010  -1

[Beefalow and the Consumer]

Here's the perl program I wrote to do this.  To run it, you supply a name of a file with a state and a population on each line, separated by a tab.  The file name is the first argument.  An optional second argument is the number of seats you want to apportion (the default is 435).

Code:

#! /usr/bin/perl

use warnings;
use strict;

if (@ARGV < 1) {
  print "Usage: app.pl population_file [number of seats]\n";
  exit(1);
}

if (@ARGV < 2) {
  $ARGV[1] = 435;
}

# read state pops
open STATEPOP, $ARGV[0] or die $!;

my %statepops;
while (my $line = <STATEPOP>) {
  chomp $line;
  my ($name, $pop) = split (/\t/, $line);
  $pop =~ s/,//g;
  $statepops{$name} = $pop;
}

my $seats = $ARGV[1] - keys(%statepops);

close (STATEPOP);

# generate app table:
my @atable;
$atable[0] = 0;
$atable[1] = 1;
for (my $i=2; $i<200; $i++) {
  push (@atable, 1/(sqrt($i*($i-1))));
}

# generate final evtable:
my %evtable;
foreach my $key (keys(%statepops)) {
  $evtable{$key} = 1;
}

for (my $i=1; $i<=$seats; $i++) {
  my $maxpri=0;
  my $state;
  foreach my $key (keys(%statepops)) {
    my $seatnum = ($evtable{$key} + 1);
    my $pop = $statepops{$key};
    my $pri = $pop * $atable[$seatnum];
    if ($pri > $maxpri) {
      $maxpri = $pri;
      $state = $key;
    }
  }
  if ($i==386) { print "-------------------\n"; }
  print "Awarding seat ", $evtable{$state}+1, " to $state.\n";
  $evtable{$state}++;
}
print "\n";

# Add senators:
foreach my $key (keys(%evtable)) {
  $evtable{$key} += 2;
}

my @sorted = sort{$evtable{$b} <=> $evtable{$a}} keys(%evtable);
foreach my $key (@sorted) {
  print $key, ": ", $evtable{$key}, "\n";
}

When I get around to it, I may create a web interface.  Although I don't currently have any place to host it (Dave?).

[Beefalow and the Consumer]

Here's another fun little use for this:  I discovered how much you would have to expand Congress for every state to have at least 2 Congressmen.  We would need 806, and they would distribute like this (with the 2 Senatorial EVs added):

California: 99
Texas: 62
New York: 56
Florida: 48
Illinois: 38
Pennsylvania: 37
Ohio: 34
Michigan: 30
New Jersey: 26
North Carolina: 25
Georgia: 25
Virginia: 22
Massachusetts: 20
Washington: 19
Indiana: 19
Missouri: 18
Tennessee: 18
Maryland: 17
Arizona: 17
Wisconsin: 17
Minnesota: 16
Louisiana: 15
Alabama: 15
Colorado: 14
Kentucky: 14
South Carolina: 14
Oregon: 12
Connecticut: 12
Oklahoma: 12
Mississippi: 10
Kansas: 10
Arkansas: 10
Iowa: 10
Nevada: 8
Utah: 8
Nebraska: 7
West Virginia: 7
New Mexico: 7
Hawaii: 6
Maine: 6
New Hampshire: 6
Idaho: 6
Montana: 5
Rhode Island: 5
North Dakota: 4
Delaware: 4
Vermont: 4
South Dakota: 4
Alaska: 4
Wyoming: 4

Wyoming, of course, gets the final seat.

[Kevinstat]

I wonder where they would put a new congressional district in Oregon.

All five congressional districts in Oregon come fairly close to each other, so the new district, if Oregon did gain a seat and there was one district that was clearly the new one, might consist of some territory from each of the existing congressional districts.  I'm not looking at any population figures here, but possible areas that could go into the new district would be Yamhill county from the 1st district, Hood River, Wasco, Sherman, and Jefferson Counties from the 2nd district, Clackamas County from the 3rd and 5th districts, eastern Multinomah County from the 3rd district, Linn County and northeastern Lane County from the 4th district, and eastern Marion county from the 5th district.  I'm sure that there will be transfers of territory between districts that don't all go to one district, however, even though if it was possible to create a plan that would withstand Judicial srutiny with 5 districts created entirely within the existing districts, with a 6th district created from the remainder of those districts.  It's not even clear that it would be mathematically possible, since the percentage of the state's population that would have to be (with little room for deviation) in each district would go from 20% to 18.33%, a decrease of only 18.33% (we're talking percentages of percentages here), and some of Oregon's congressional districts might lose more than a sixth of it's resident population (according to the 2010 census and relative to the state's resident population) over the course of this decade, although now that I am looking at the changes in population of the 1990's districts that seems very unlikely.

Oregon's 1st and 5th districts currently look kind of messy, particularly the latter.  I think regardless of whether or not Oregon gains a 6th seat, there should be one congressional district entirely in Multnomah county and including all of Portland and it's immeidate suburbs.  It would be preferable if the remainder were all in one district and be a clear section of the county, but that might not be possible without having the line go through Portland or it's immdediate suburbs.  There is more territory in Multnomah County east of Portland than west of it, but that doesn't necessarily translate into more people and eastern Multnomah County has been in the 3rd district since at least the 1992 elections.  I'm having a lot of interesting thoughts about the possibilities for the new Oregon districts, but it's way to early to have any guess about what the new districts will look like, unless Oregon doesn't gain a 6th seat in which case it's likely that the changes will be fairly minor.

Sincerely,

Kevin Lamoreau

[Kevinstat]

Apparently the version of Dave's map generator that I am using that has PR available for display has the number of votes displayed for Maine and Nebraska hardcoded to their EV's.  That's how they gained two seats each.

Let me give it a try:

You must be logged in to read this quote.

[Kevinstat]

Okay, trying again:

[A18]

Why did you give the Atlantic Ocean a representative?

[Kevinstat]

Maybe if I change the year in the code to 1968, before Maine adopted it's present system of choosing electors...



Yes, it works!  It doesn't work with 2012 though.  And wait, no Puerto Rico in my 1968 try.  I give up.
Phillip,

That's the District of Columbia.  That box appears in that location in all of the calculator-generated (if an election in or after 1964 is selected) and prediction maps.  That's a funny thought, though.

[jimrtex]

Montana will not get another seat.

Montana came close to getting another seat this time patrick.

Montana was even closer in 1990 when they lost the 2nd seat and sued over the apportionment method (presidential trivia: the lawyers who argued before the USSC were Racicot and Starr).

During the 1990's Montana gained 12.9%, while the country as whole gained 13.2%, meaning that Montana was losing relative share at about -0.3%.  This might not be hard to make up, but it is not guaranteed by any means.

[jimrtex]

The last few seats I apportioned were:
432 - CA 55
433 - PA 18
434 - MN 8
435 - MI 15

The next in line were
NY 28
IL 19
AL 7

Then after a gap
CA 56
MD 9
OH 17
TX 36

The curiousity is that the states at the edge are all losing seats (with California bracketing them).

In other words, there are 11 clear gains and 8 clear losses, and then 6 fighting to decide which 3 stay get to stay.

The number of representatives for California will tend to be volatile.  With over the difference between 53/435 and 54/435 is less than 2%,
meaning that California can on average gain a seat with a rate of increase that is less than 2% greater than the national average.  On the other hand, to increase from 2 to 3 representatives, a state will have to increase 73% faster than the national average (a state with a population a bit larger than Montana, or with a bit less than Nebraska, would still get 2 representatives).

With about 1/8 of the population, California will be roughly every 8th step on the priority list, with each priority value about 2% less than the previous.  Whether there are 10 other states between consecutive California seats of 6 states is pretty much random.  Other states may occasionally be near the edge (North Carolina and Utah in 2000), but by 2010 they should be well above that level.  California will always be nearby.  Sometimes to get an extra seat as in 2000, sometimes to lose a seat.

[Filuwaúrdjan]

What if the Apportionment was based on voters, not population?

[jimrtex]

What if the Apportionment was based on voters, not population?

The following is based on the 2000 election (just because they are more complete).   14 states gain, with big pickups in the midwest.   Losses in 6 states, especially concentrated in California and Texas with their large immigrant populations.


Alabama              7
Alaska               1
Arizona              6
Arkansas             4
California          45
Colorado             7
Connecticut          6
Delaware             1
District of Columbia 1
Florida             25
Georgia             11
Hawaii               2
Idaho                2
Illinois            20
Indiana              9
Iowa                 5
Kansas               4
Kentucky             6
Louisiana            7
Maine                3
Maryland             8
Massachusetts       11
Michigan            18
Minnesota           10
Mississippi          4
Missouri            10
Montana              2
Nebraska             3
Nevada               3
New Hampshire        2
New Jersey          13
New Mexico           3
New York            28
North Carolina      12
North Dakota         1
Ohio                19
Oklahoma             5
Oregon               6
Pennsylvania        20
Rhode Island         2
South Carolina       6
South Dakota         1
Tennessee            9
Texas               27
Utah                 3
Vermont              1
Virginia            11
Washington          10
West Virginia        3
Wisconsin           11
Wyoming              1

[jimrtex]

And based on the 2000 election this would be the apportionment of Congoress and the House of Republicans

(Numbers don't add to 435 due to use of rounding, rather a strict priority list.)


Alabama              6   8
Alaska               1   2
Arizona              6   7
Arkansas             4   4
California          50  39
Colorado             6   8
Connecticut          7   5
Delaware             2   1
District of Columbia 2   1
Florida             25  25
Georgia             10  12
Hawaii               2   1
Idaho                1   3
Illinois            22  17
Indiana              8  11
Iowa                 5   6
Kansas               3   5
Kentucky             5   8
Louisiana            7   8
Maine                3   3
Maryland            10   7
Massachusetts       14   8
Michigan            19  17
Minnesota           10  10
Mississippi          3   5
Missouri            10  10
Montana              1   2
Nebraska             2   4
Nevada               2   3
New Hampshire        2   2
New Jersey          15  11
New Mexico           3   3
New York            35  21
North Carolina      11  14
North Dakota         1   2
Ohio                19  20
Oklahoma             4   6
Oregon               6   6
Pennsylvania        21  20
Rhode Island         2   1
South Carolina       5   7
South Dakota         1   2
Tennessee            8   9
Texas               21  33
Utah                 2   4
Vermont              1   1
Virginia            10  12
Washington          11  10
West Virginia        3   3
Wisconsin           11  11
Wyoming              1   1