Colorado: another nail in the elctral collg coffin

[jimrtex]

Not the procedure for making a "final determination of any contest or controversy," which is what 3 U.S.C. §5 is about.  You're talking about changing the procedure for filing such a contest.

3 U.S.C. §5 says that if there is a procedure for making a final determination as to who the presidential electors are; AND if that procedure is in place prior to the election; AND if the procedure as defined at the time of the election was used to make a determination at least 6 days before the meeting of the electors; THEN that determination will be conclusive.

If no contest is filed, no controversy exists for the Supreme Court to make a final determination of.  Changing the filing deadline changes the procedure.

Colorado law provides that votes for candidates will be counted before votes for ballot issues.  If the votes for presidential tickets are canvassed and it is determined that according to the law that exists at the time that 9 Bush electors have been elected, on what basis could that result be contested?  That there is the possiblity that the legislature in the future may adopt a retroactive interpretation of the votes cast?

[jimrtex]

Under current Colorado law, if a voter places an X in the box next to a candidate's name he knows that if his preferred candidate receives the most votes. that all of the candidate's presidential electors will be chosen.

That is, when a voter casts his vote, he knows it will have a particular effect beyond the trivial increase in the opacity of the ballot paper, and that this may influence who he votes for.  While a voter is not guaranteed a particular outcome, he should be able to know how that outcome is determined.

The method by which electors are chosen may influence who contests the election.  For example, a candidate may receive an EV in Colorado with as little as 5.6% of the vote.  The filing deadline for independent candidates in Colorado is 120 days before the election, which is prior to the deadline for filing initiatives.  The number of signatures required for ballot access may be reasonable for a winner-take-all statewide election.  But it may not be reasonable for a proportional election.   Nationwide, there were roughly 200,000 votes cast per EV.  If a candidate can receive an EV with a vote share equal to about 0.5 of this (100,000) what sort of petition requirement is reasonable.  If you only need 100,000 votes to be elected, is it reasonable to require 50,000 signatures to even get on the ballot?  I don't think so.

Don't all of these factors infringe on the right to vote?

[jimrtex]

If the proposed Colorado method of allocating electors had been in place nationwide in 2000, the result would have been Gore 269, Bush 263, and Nader 6.  With no candidate having a majority, the election of President and Vice President would have been turned over to House and Senate, respectively.

The popular vote in Florida would determine whether Gore or Bush prevailed in a 13-12 split.  A Gore win in Florida would have meant a national Gore victory.

But there would be a total of 13 States where a shift of less than 1% of the total vote would have changed the allocation of EVs for that State.  In the most favorable case, Bush would have won 272-259-7.

The closeness is not all that surprising, considering that the national popular vote difference was equivalent to about 2 EV.

In 2000, there were 5 States where Gore defeated Bush by less than 1%, Iowa, New Mexico, Oregon, and Wisconsin.  None of these were particularly controversial once it was realized that if Gore won in Florida, Bush would have had to win in the largest 3 to be elected.

Under the Colorado method, we would have the potential of more than 1/4 of the States being as controversial as Florida.

Arizona: If Gore lost 0.92% of the total vote, a 4-4 split would have switched to a 3-5 split (If the 2000 popular vote were repeated, Bush would win both additional EV in Florida).

California: If Bush gained 0.02% of the vote, he would have gained an addition EV (30-22-2 to 29-23-2).  In a state Bush lost by 1.3 million, a mere 1677 votes would tip the results.

Florida: A Gore gain of 0.01% would change the result from 12-13 to 13-12.

Georgia: A Gore loss of 0.68% would have switched a 6-7 result to 5-8.  That loss is only slightly more than the number of votes that Nader received as write-in candidate.

Hawaii: An additional 137 Bush votes (0.04% of the total) would have produced a 2-2 split.

Illinois: An additional 0.08% for Nader would have given him an EV at the expense of Gore.

Iowa: An additional 0.31% would have given Bush an additional EV.

Minnesota: A Bush loss of 0.50% would have given Nader an EV at the expense of Bush.

New Mexico: A Bush gain of 0.06% would have given Bush a 3-2 victory.

Ohio: An additonal 1520 votes (0.03%) would have given Bush an additional EV at the expense of Nader.   Bush carried Ohio by 170,000 votes.

Oregon: A Bush gain of 0.44% would have given Bush a 4-3 victory.

Washington: A Nader Gain of 0.41% and a Gore loss of 0.13% would have switched an EV from Gore to Nader.  Washington has liberal laws with respect to voting by mail, and counts votes that are postmarked on election day.  In 2000, the count in the Senate race stretched on for weeks, with counties eventually refusing to disclose how many ballots remained uncounted.

Wisconsin: A Bush gain of 0.22% would have secured 6-5 victory.

[Angel of Death]

It is generally agreed among academics that the fairest (least size-biased) way of handling fractions is the Webster or Sainte-Laguë method, which works by adjusting the quota such that the sum of the quotients rounded to their closest integer would equal the target total. Applying this to the 2000 presidential election, we get the following:

Alaska
Bush: 2
Gore: 1

Alabama
Bush: 5
Gore: 4

Arkansas
Bush: 3
Gore: 3

Arizona
Bush: 4
Gore: 4

California
Bush: 23
Gore: 29
Nader: 2

Colorado
Bush: 4
Gore: 4

Connecticut
Bush: 3
Gore: 5

District of Columbia
Gore: 3

Delaware
Bush: 1
Gore: 2

Florida
Bush: 13
Gore: 12

Georgia
Bush: 7
Gore: 6

Hawaii
Bush: 2
Gore: 2

Iowa
Bush: 3
Gore: 4

Idaho
Bush: 3
Gore: 1

Illinois
Bush: 10
Gore: 12

Indiana
Bush: 7
Gore: 5

Kansas
Bush: 4
Gore: 2

Kentucky
Bush: 5
Gore: 3

Louisiana
Bush: 5
Gore: 4

Massachusetts
Bush: 4
Gore: 7
Nader: 1

Maryland
Bush: 4
Gore: 6

Maine
Bush: 2
Gore: 2

Michigan
Bush: 9
Gore: 9

Minnesota
Bush: 4
Gore: 5
Nader: 1

Missouri
Bush: 6
Gore: 5

Mississippi
Bush: 4
Gore: 3

Montana
Bush: 2
Gore: 1

North Carolina
Bush: 8
Gore: 6

North Dakota
Bush: 2
Gore: 1

Nebraska
Bush: 3
Gore: 2

New Hampshire
Bush: 2
Gore: 2

New Jersey
Bush: 6
Gore: 9

New Mexico
Bush: 2
Gore: 3

Nevada
Bush: 2
Gore: 2

New York
Bush: 12
Gore: 20
Nader: 1

Ohio
Bush: 10
Gore: 10
Nader: 1

Oklahoma
Bush: 5
Gore: 3

Oregon
Bush: 3
Gore: 4

Pennsylvania
Bush: 11
Gore: 12

Rhode Island
Bush: 1
Gore: 3

South Carolina
Bush: 5
Gore: 3

South Dakota
Bush: 2
Gore: 1

Tennessee
Bush: 6
Gore: 5

Texas
Bush: 19
Gore: 12
Nader: 1

Utah
Bush: 4
Gore: 1

Virginia
Bush: 7
Gore: 6

Vermont
Bush: 1
Gore: 2

Washington
Bush: 5
Gore: 6

Wisconsin
Bush: 5
Gore: 6

West Virginia
Bush: 3
Gore: 2

Wyoming
Bush: 2
Gore: 1

Total: 538
Bush: 265
Gore: 266
Nader: 7

[jimrtex]

It is generally agreed among academics that the fairest (least size-biased) way of handling fractions is the Webster or Sainte-Laguë method, which works by adjusting the quota such that the sum of the quotients rounded to their closest integer would equal the target total. Applying this to the 2000 presidential election, we get the following:

Total: 538
Bush: 265
Gore: 266
Nader: 7

It would be interesting to know why the Colorado method was chosen.  If the 3rd party is shut out, the Colorado method is equivalent to Saint-Lague after the votes of the 3rd and lower candidates is given to the leading candidate.  That is, pure Saint-Lague disregards the votes of the trailing candidates, while the Colorado method gives them to the leading candidate before distributing the EV.

Nationwide, pure Saint-Lague would favor Bush (under a nationwide application of the Colorado method, it would be Gore 269, Bush 263, Nader 6).

But ironically, Colorado was the one state where this would benefit Bush at the expense of Gore.  So perhaps the Colorado method was chosen because it produced an EV victory for Bush.  Instead of having to explain how an 8.4% popular vote win translates into a 4:4 EV tie, it can be shown as a demonstration of one candidate being rewarded for his popular vote victory, but not to the exclusion of the losing candidate.

The 6 states where pure Saint-Lague produces a different result from the Colorado method are:

Hawaii, Illinois, Michigan, and California: Bush picks up a Gore EV under pure Saint-Lague.  In all but California, this is mostly treating the Nader vote as if it were for the leader, Gore.  In California, it is more treating the 4th and beyond (Browne, Buchanan, etc.) as being for Gore.

Colorado: Gore picks up a Bush EV under pure Saint-Lague.  Here, Nader's votes were treated as being for Bush.

Minnesota: Gore picks up a Bush EV under pure Saint-Lague.  Minnesota was the one state under the Colorado method where all 3 candidates rounded up to a higher number of EV.  The unrounded share of the 10 EV was Gore 4.79, Bush 4.55, Nader 0.52, others 0.14).  After rounding it was Gore 5, Bush 5, Nader 1, or one extra EV being allocated.  Under the Colorado rules, the extra EV is always taken from the last place candidate, in this case Nader.

[Angel of Death]

It is perhaps noteworthy enough to mention that the current congressional apportionment is also the exact result if it was done by Webster even without enforcing a minimal 1 seat per state.

[jimrtex]

It is perhaps noteworthy enough to mention that the current congressional apportionment is also the exact result if it was done by Webster even without enforcing a minimal 1 seat per state.

Some may infer from this that the two methods will always produce the same results, or that the seat-by-seat apportionment order would be the same.

For example, if Webster's method were used, then there wouldn't have been an issue of overseas Mormon missionaries not being counted in Utah's apportionment population.  Alternatively, if Utah had won their law suit, but then Congress switched from equal proportions to Webster's, Utah would have again lost their 4th representative.

The most significant difference is for States with low population.  Rhode Island current priority rank for their 2nd seat is 385 (even if House were reduced in size from 435 to 385, they would keep their 2 representatives.  But if Webster's method were used, then they would be ranked around 400.   A State with 914,000 would have 2 representatives under equal proportions.   They would need 967,000 under Webster's.

[Angel of Death]

In other words, Huntington-Hill (the currently used method) is biased in favor of smaller states.

[muon2]

In other words, Huntington-Hill (the currently used method) is biased in favor of smaller states.

It uses a geometric mean - the square root of the product of two consecutive numbers. This value is farthest from the average for the smallest numbers. That's what gives extra weight to smaller states.

[jimrtex]

In other words, Huntington-Hill (the currently used method) is biased in favor of smaller states.

In other other words, it provides an exquisite balance between the principle that States should be proportionately represented, and the principle that similar-sized groups of people should elect a representative.

[freedomburns]

Does anyone have info on the likelihood of Amendment 36 passing?  Are there any polls out?  Does it need >50% to pass?

[Swing low, sweet chariot. Comin' for to carry me home.]

Ever wonder why this idea isn't being attempted in California?

[WMS]

Ever wonder why this idea isn't being attempted in California?

Or Texas?

[WMS]

It is generally agreed among academics that the fairest (least size-biased) way of handling fractions is the Webster or Sainte-Laguë method, which works by adjusting the quota such that the sum of the quotients rounded to their closest integer would equal the target total. Applying this to the 2000 presidential election, we get the following:

[snip for space]

Total: 538
Bush: 265
Gore: 266
Nader: 7

What are the results if, say, a 5% threshold is applied in each state? How badly does Nader suffer from it?

And this is a superb thread! Kudos to all those who posted!

[jimrtex]

Ever wonder why this idea isn't being attempted in California?

Or Texas?

There is no initiative in Texas, so even if a rich Californian wanted to fund such an effort, he couldn't.

[jimrtex]

It is generally agreed among academics that the fairest (least size-biased) way of handling fractions is the Webster or Sainte-Laguë method, which works by adjusting the quota such that the sum of the quotients rounded to their closest integer would equal the target total. Applying this to the 2000 presidential election, we get the following:

[snip for space]

Total: 538
Bush: 265
Gore: 266
Nader: 7

What are the results if, say, a 5% threshold is applied in each state? How badly does Nader suffer from it?

And this is a superb thread! Kudos to all those who posted!

Under the Colorado proposal with a 5% threshold, Nader would lose EV in California(2), New York, Ohio, and Texas, keeping only his one EV from Massachusetts.   Under pure St-Lague he would also keep his one EV from Minnesota.

Under the Colorado proposal, there is effectively a threshold of 1/2 the total number of EV, so any state with 10 EV or fewer has a threshold of 5% or greater.  Under pure St-Lague, the threshold is is 1/2 of the total number of EV + 0.5, with the percentage figured on the 3-party vote, rather than the total vote.  In a state with 10 EV, the threshold would be 5% under the Coloard proposal, and about 4.76% under pure St-Lague.

Under the Colorado proposal, the winner in a state would gain any votes that an unsuccessful candidate did not receive, so Nader's lost votes would be split 3 for Gore (New York and California) and 2 for Bush (Texas and Ohio).  Under St-Lague with a threshold, the extra EV would be allocated on the basis of the Bush-Gore vote.  Bush and Gore would split the 2 California EV, Gore would gain the New York EV, and Bush would gain the Texas and Ohio EV.

[Angel of Death]

In other words, Huntington-Hill (the currently used method) is biased in favor of smaller states.

In other other words, it provides an exquisite balance between the principle that States should be proportionately represented, and the principle that similar-sized groups of people should elect a representative.

I disagree; the latter principle is supposed to have been applied through the existence of the Senate. Anyway, the difference in practice can hardly be called mind-shattering.

[Angel of Death]

Case in point: No difference between Webster and Huntington-Hill with 1990 as well.

[jimrtex]

In other words, Huntington-Hill (the currently used method) is biased in favor of smaller states.

In other other words, it provides an exquisite balance between the principle that States should be proportionately represented, and the principle that similar-sized groups of people should elect a representative.

I disagree; the latter principle is supposed to have been applied through the existence of the Senate. Anyway, the difference in practice can hardly be called mind-shattering.

If your goal is to have similarly-sized groups of people elect a representative, regardless of which state they live in, then you don't want to have a single representative have to represent 50% more than the typical representative does, while a pair of representatives at worst each only have to represent 75% of a typical representative does.  It would be better to have the worst case to have 33% extra constituents or 33% fewer constituents than is average.

But you might not want a state with an entitlement of only 1.33 representatives to have 2 representatives.

So you choices of divisors are:

(n+(n-1))/2  Better apportionment of represntative per population, but greater deviation in population per representative.

(2*n*(n-1))/(n+(n-1)) Less deviation in population per representative but greater deviation in representative per population.

sqrt(n*(n-1)) Just right - does moderately well on both tests.

The Senate does the opposite of providing equal population per representative.  It provides the same number of Senators regardless of the population.

[jimrtex]

Case in point: No difference between Webster and Huntington-Hill with 1990 as well.

Are you sure that Oklahoma would not have lost their 6th representative and Massachusetts have kept their 11th if Webster had been used?  See footnote 13 in DEPARTMENT OF COMMERCE v. MONTANA, 503 U.S. 442 (1992).

[Angel of Death]

In other words, Huntington-Hill (the currently used method) is biased in favor of smaller states.

In other other words, it provides an exquisite balance between the principle that States should be proportionately represented, and the principle that similar-sized groups of people should elect a representative.

I disagree; the latter principle is supposed to have been applied through the existence of the Senate. Anyway, the difference in practice can hardly be called mind-shattering.

I think the words "States should be proportionately represented" put me on the wrong track there and that my response to it above should be discarded. Anyway, Webster can be succinctly rationalized in the following way: What quota you use doesn't matter per se, as long as in the end it all adds up if the same value is used constantly. The original quota itself only works if you would have been allowed to work with fractions (as it would by definition) and shouldn't in any way be considered "sacred" beyond that.

Case in point: No difference between Webster and Huntington-Hill with 1990 as well.

Are you sure that Oklahoma would not have lost their 6th representative and Massachusetts have kept their 11th if Webster had been used?  See footnote 13 in DEPARTMENT OF COMMERCE v. MONTANA, 503 U.S. 442 (1992).

I'm using the values from http://www.census.gov/population/censusdata/apportionment/table-a.pdf

[jimrtex]

Case in point: No difference between Webster and Huntington-Hill with 1990 as well.

Are you sure that Oklahoma would not have lost their 6th representative and Massachusetts have kept their 11th if Webster had been used?  See footnote 13 in DEPARTMENT OF COMMERCE v. MONTANA, 503 U.S. 442 (1992).

I'm using the values from http://www.census.gov/population/censusdata/apportionment/table-a.pdf

Massachusetts: 6,029,051
Oklahoma: 3,157,604

MA/10.5 = 574,195
OK/5.5 = 574,110

MA/sqrt(10*11) = 574,847
OK/sqrt(5*6) = 576,497

[Angel of Death]

Good call. Apparently, the program I used couldn't handle close calls.

[WMS]

Ever wonder why this idea isn't being attempted in California?

Or Texas?

There is no initiative in Texas, so even if a rich Californian wanted to fund such an effort, he couldn't.

Oh darn, a technicality. I was just having fun with the fact that both parties want to try out ideas to split state electoral votes...













...in states that the other party won.

[WMS]

It is generally agreed among academics that the fairest (least size-biased) way of handling fractions is the Webster or Sainte-Laguë method, which works by adjusting the quota such that the sum of the quotients rounded to their closest integer would equal the target total. Applying this to the 2000 presidential election, we get the following:

[snip for space]

Total: 538
Bush: 265
Gore: 266
Nader: 7

What are the results if, say, a 5% threshold is applied in each state? How badly does Nader suffer from it?

And this is a superb thread! Kudos to all those who posted!

Under the Colorado proposal with a 5% threshold, Nader would lose EV in California(2), New York, Ohio, and Texas, keeping only his one EV from Massachusetts.   Under pure St-Lague he would also keep his one EV from Minnesota.

Under the Colorado proposal, there is effectively a threshold of 1/2 the total number of EV, so any state with 10 EV or fewer has a threshold of 5% or greater.  Under pure St-Lague, the threshold is is 1/2 of the total number of EV + 0.5, with the percentage figured on the 3-party vote, rather than the total vote.  In a state with 10 EV, the threshold would be 5% under the Coloard proposal, and about 4.76% under pure St-Lague.

Under the Colorado proposal, the winner in a state would gain any votes that an unsuccessful candidate did not receive, so Nader's lost votes would be split 3 for Gore (New York and California) and 2 for Bush (Texas and Ohio).  Under St-Lague with a threshold, the extra EV would be allocated on the basis of the Bush-Gore vote.  Bush and Gore would split the 2 California EV, Gore would gain the New York EV, and Bush would gain the Texas and Ohio EV.

Thank you very much! I wonder how well Perot would've done in 1992 and 1996...?