Category Archives: Election 2004

Election Week 2004

The plan for Election Week: Starting Monday, the main election result database will be off-line. Site members with Individual, Group, or Student licenses will continue to have access. This is necessary in order to reduce server load and conserve on bandwidth. I apologize for any inconvenience that this may cause.
Election predictions will close at 11pm EST on Monday. The polls close on the Mock Election at 5pm EST on Tuesday.

On Tuesday, I plan to update the time-line once every 30 minutes at about 10 past and 40 past each hour. State and county-level results will be posted once states reach ~ 90% precinct reporting, but not starting until around 10pm. On Wednesday morning, preliminary state and county data will be on-line. County maps will also be available. CD maps and data are only availble for those nice states with boundaries that do not cross county/town lines.

The 2004 election results will be available linked from the home page and the 2004 index page ( the national link page). Note that non-site-members will see ads with these pages. This is temporary during the high-bandwidth days. On Tuesday, the 2004 menu options will also appear in the main database for members to access.

Maine Trend Analysis

The latest Survey USA Maine Poll shows a significant gap in support for George W. Bush and John Kerry between Maine’s two congressional districts: Kerry +4 in ME1, Bush +5 in ME2, a nine point gap (caveat – the poll lists “Northern Maine and “Southern Maine” and it is unknown how well these line up with actual congressional district boundaries). The previous entry showed that in 2000, the difference between the results for Gore and Bush in the current districts is 4.4 percentage points.

The table below shows the trend data for the election results from 1980 through 2000 in the current districts:

Year ME-D CD1-D CD2-D ME-R CD1-R CD2-R
1980 42.3% 41.9% 42.6% 45.6% 45.0% 46.2%
1984 38.8% 39.7% 37.9% 60.8% 59.9% 61.7%
1988 43.9% 42.9% 44.9% 55.3% 56.3% 54.3%
1992 38.8% 39.8% 37.8% 30.4% 32.3% 28.5%
1996 51.6% 51.7% 51.5% 30.8% 32.2% 29.3%
2000 49.1% 50.2% 47.9% 44.0% 42.9% 45.1%

The next table shows the difference between each congressional district and the statewide total by year:

Year CD1-D CD2-D CD1-R CD2-R
1980 -0.3% 0.3% -0.6% 0.5%
1984 0.9% -0.9% -0.9% 0.9%
1988 -1.0% 1.0% 1.0% -1.0%
1992 1.0% -1.0% 1.9% -1.9%
1996 0.1% -0.1% 1.4% -1.4%
1996 1.1% -1.1% -1.0% 1.1%

The trend graphs for each district are shown below:

Maine Trend Graph: Congressional District 1 vs. Statewide
CD1 Graph

Maine Trend Graph: Congressional District 2 vs. Statewide
CD2 Graph

Although its difficult to conclude any solid trend in Maine from these graphs (the strength of Ross Perot in 1992 and 1996 and Nader in 2000 certainly masks some of the Republican/Democratic trends), there does appear to be a slight trend in voting behavior over the last decade – increasing the difference between Republican and Democratic voters between the two districts (CD2 becoming more Republican and CD1 more Democratic). While I don’t believe it to be nine percentage points as the SurveyUSA poll suggests, I think it bears watching and could result in a split Electoral Vote allocation if the statewide result is close.

The Potential for Split Electors in Maine

There has been some discussion with regard to the prospects this year for Maine to split its allocation of electoral votes. Maine is one of two states (Nebraska is the other) that chooses its electors through the congressional district method. That is, the two at-large electors pledged to candidate whom receives the pluraility of popular votes statewide are chosen (representing the state’s two senators) and the elector pledged to the candidate whom receives the plurality of votes within each congressional district is chosen (representing each district’s representative). Since this method was instituted in 1972, Maine has never split its presidential electors.

The prospects for a split verdict in 2004 gained a bit of interest with this recent Survey USA Poll that included the following statement: “Note: Maine is not a ‘Winner-take-all’ electoral state. If Bush Wins ME CD #2, which we label here ‘Northern Maine’, and where he is now tied, Bush would receive 1 of the state’s 4 electoral votes.” The poll shows 135 respondents for Kerry and 135 for Bush in “Northern Maine”. Unfortunately, from the poll, we can not tell what geographic areas constitute “Northern Maine”.

Some numberical analysis: In 2000, the second congressional district was won by Gore by 5,660 votes (or 1.87%). The table below displays the data for Maine in 2000 by congressional district.

CD Total Vote Margin %Mar %Dem %Rep %Grn %Oth Dem Rep Grn Oth
1 348,951 27,675 7.93% 50.52% 42.59% 5.82% 1.07% 176,293 148,618 20,297 3,743
2 302,866 5,660 1.87% 47.43% 45.56% 5.56% 1.45% 143,658 137,998 16,830 4,380

From the data, a change of only 2,831 votes (about 0.9%) from Gore to Bush in the second district would have resulted in Bush winning one electoral vote from Maine.

In 2003, Maine performed its decennial redistricting, the result of which had a net result of shifting a significant number of Democrats from the first district into the second district. The court-ordered boundaries moved seven Kennebec County towns from the first district to the second (Benton, Clinton, Fayette, Litchfield, Oakland, Waterville, and Winslow). In 2000, these towns gave Gore a margin of 3,344 votes. The changes also moved three towns from the second to the first district (Albion, China, and Monmouth). These three towns gave Bush a margin of 78 votes in 2000. Overall, with the 2004 congressional district boundaries, Gore gains 3,422 votes, about 1.1 percentage points. A detailed town results table is available to members. The table below shows the 2000 results using the 2004 congressional district boundaries:

CD Total Vote Margin %Mar %Dem %Rep %Grn %Oth Dem Rep Grn Oth
1 335,467 24,253 7.23% 50.16% 42.93% 5.84% 1.07% 168,266 144,013 19,599 3,589
2 316,350 9,082 2.87% 47.95% 45.08% 5.54% 1.43% 151,685 142,603 17,528 4,534

As shown in the table above, the redistricting moved the margin of the two districts closer by about 1.8 percentage points.

A second take on this is to look at the voter registration. The table below shows the January 2004 voter enrollment from the Maine Bureau of Corporations, Elections, and Comissions

CD Total Voters %Dem %Rep %Une %Grn Dem Rep Une Grn
1 492,995 30.6% 28.8% 38.4% 2.2% 151,025 141,937 189,252 10,781
2 464,490 31.6% 28.6% 38.0% 1.8% 146,806 132,790 176,669 8,225
Net Change 2000 – 2004
1 -3,226 -0.8% -0.8% 0.6% 1.2% -5,098 -4,689 1,904 6,044
2 13,522 0.3% -0.6% -0.2% 0.9% 5,524 1,188 4,318 4,219

From the table, the approximate 1.0 percentage point net change between Democrats and Republicans in congressional district 2 (from 2% Democratic advantage to 3% Democratic advantage) mirrors the 1.0% delta margin change in the 2000 election result using the 2000 vs. 2004 Congressional District boundaries (1.87% to 2.87% as shown in the tables above). However, the very large number of Independent voters overwhelms this small difference between Democratic and Republican registration. The data also show that the Independents in the first district more heavily favored Gore over Bush than their unenrolled counterparts in the second district (assuming that the voter turnout rates amongst the different parties are similar across districts)

Although the prospects for a split verdict in Maine are possible, the net effect of redistricting has reduced the probability. I expect that Bush would have to increase his national percentage by 3-5 percentage points to carry Maine’s second district.

Allocating Electors Proportionally

This is a follow-up post to the description of the current ballot initiative in Colorado posted yesterday. The voters on November 2, 2004 will act in place of the state legislature and vote on an initiative to change the manner in which the state chooses Presidential Electors from the winner-take-all method to a new system of allocating electors amongst the candidates in a manner approximately proportional to their respective popular vote percentages (there is additional complexity in the algorithm that deals with rounding errors).

To my knowledge, this methodology for choosing Presidential Electors has no precedent. There are several pros and cons to this method:

  • Pro: It could eliminate the need for a controversial recount for close state-wide elections (e.g. the Florida result in 2000 would not likely have been contested because the result would have split 13 Electoral Votes for Bush and 12 for Gore). However, a Con of this approach is the possibility of many recounts nationwide in multiple states over a few electoral votes may occur if the result is close nation-wide.
  • Pro: It allocates the electors in a manner that more closely represents the popular vote. This has the appeal of being fair.
  • Pro: The method reduces the “wasted vote” syndrome. Today, if a state is not competitive, those citizens casting votes for the minority party in that state are not contributing electorally nation-wide (e.g. Democrats in Utah or Republicans in Maryland). Additionally, minor party candidates have a much greater chance of being awarded Electoral Votes. However, a Con of this is the increased chance of the election being decided by the House of Representatives (due to no candidate receiving a majority of electoral votes).
  • Pro: With a proportional system nationwide, there is greater likelihood of candidates visiting and campaigning in states that are not currently competitive. A serious Con: in the case of Colorado, however, is that since this measure only goes into effect in one state, it effectively removes the battleground status of the state. Instead of competing for 9 EVs, the Kerry and Bush would be campaigning for one.
  • Pro: A nationwide proportional system reduces (but does not eliminate) the probability of a split verdict between the popular vote and the electoral vote (as was witnessed in 2000).

Another point that could be argued either for or against the proposed change is that winner-take-all method (usually) exaggerates the margin of victory. Typically, a landslide electoral victory can be achieved with a simple majority in the popular vote, giving the impression of a mandate and national unity. For example, in 1984 Ronald Reagan won 58.8% of the popular vote, but 97.6% of the Electoral Vote. Likewise Franklin Roosevelt won 57.4% of the popular vote in 1932, but 88.9% of the Electoral Vote.

Furthermore, the manner in which the initiative dictates that electors be added or subtracted from candidates when there are rounding errors should certainly be debated with regard to fairness. Its a very different algorithm than the method used to assign congressional seats.

Overall, I think the idea of choosing Presidential Electors proportional to a candidates’ performance in the popular vote is interesting, but for Colorado to enact the method on its own will essentially eliminate any influence the state currently has on the campaign.

Colorado and Proportional Allocation

Recently, the Colorado Secretary of State Donetta Davidson recently certified that the ballot initiative, Amendment 36 to the Colorado Constitution, to change the method of choosing Presidential Electors from the “Winner-take-all” system to a proportional system based on popular vote had a sufficient number of signatures to be sent to the voters on November 2, 2004. The proposed amendment includes a subsection that specifically states that the new method of allocating electors applies to the 2004 Presidential Election.

Currently, Colorado chooses its presidential electors in the same manner as 48 other states and the District of Columbia – using the winner-take-all method.

from the Colorado State Constitution:
Schedule Section 20. Presidential electors after 1876. The general assembly
shall provide that after the year eighteen hundred and seventy-six the
electors of the electoral college shall be chosen by direct vote of the people.

This is actually a rather vague description. Generally, in a Winner-Take-All popular-vote method, a slate of Electors is pledged to each Presidential/ Vice-Presidential ticket (the Electors’ names may or may not appear on the ballot). The slate of Electors pledged to the ticket having received the plurality of votes state-wide are chosen.


Methodology summary from subsections (3) and (4): Electors are allocated in whole numbers only. The percentage of vote cast for each ticket is multiplied by the total number of available electors and rounded to the nearest whole number. If the sum of the total electors allocated is greater than the number available to be appointed, then the total electoral votes for the candidate having received the fewest number of ballots (that received at least one electoral vote) is reduced by one. If the sum of the total electors allocated is less than the total number of available electors, the Presidential Ticket receiving the greatest number of ballots is granted the remaining unallocated electors. Additional clarifications are included in the section with regard to ties, recounts, etc.

If this measure were in effect for the 2000 election, the results would be:

Candidate Total Votes Vote % EV*% Rounded EV
Bush 883,745 50.8% 4.06 4
Gore 738,227 42.4% 3.39 3
Nader 91,434 5.3% 0.42 0

Since the total of allocated electors is seven and the total available is eight, one additional elector is awarded to Bush, for a final tally of Bush 5, Gore 3. Under this hypothetical scenario, Gore wins the election with 271 Electoral Votes.

The Colorado Consitution reserves to the people the right to act in place of the state legislature. This clause gives the people of Colorado the ability to decide how electors are chosen as Article II, Section I of the United States Constitution states Each State shall appoint, in such Manner as the Legislature thereof may direct, a Number of Electors …

The initiative goes further, including subsection 5 (methodology for recounts of votes for each the ballot initiative and Presidential Electors), subsection 6 (election certifications of each the ballot initiative and Presidential Electors), subsection 7 (granting the Secretary of State the power to determine, by lot, which Presidential Electors are nominated from Presidential Tickets that qualify for at least on electoral vote), subsection 8 (Supreme Court original jurisdiction for the adjudication of all contests concerning Presidential Electors), subsection 9 (effectivity date of November 3, 2004), subsection 10 (“This section shall be liberally construed to achieve popular proportional allocation of Presidential Electors at the 2004 General Election”) and subsection 11 (stating that the “general Assembly may enact legislation to change them anner of selecting Presidential Electors or any of the Procedures related thereto” – seemingly granting the legilature the ability to undo this amendment).

Florida and Ohio

This is a post relating to my comment to wormwood’s post with regard to the possiblity of a victory for Bush without Ohio or Florida. His assertion is that it is unlikely that a Republican candidate can win any midwestern states (excluding IN, MO) and simultaneously lose Ohio. Below is a trend graph that highlights the recent electoral history of the result for the Republican Presidential Candidate (the percentages are relative to the result for each Republican nationally in the respective election years)

trend graph

This graph includes results for the midwestern states of Minnesota, Iowa, Wisconsin, Missouri, Ohio plus Florida. The rank order of these midwestern states is idential between 1996 and 2000 – and almost parallel. From these data, wormwood’s assumption that other midwestern states will not likely “cross-over” to Bush without Ohio is reasonable.

Note the opposite trend in Florida. If the voting trend shown in the graph continues into 2004, it appears that Bush will lose Florida before he will lose Ohio. Without Florida, Bush must pick up Wisconsin, Iowa, and New Mexico or Oregon, while keeping New Hampshire and West Virginia.

The Prospects for a Tie

With the current projections for a close election between George W. Bush and John Kerry on November 2, there are a large number of reasonable scenarios that result in no electoral majority (i.e. a tie). In the event of no electoral majority, the U.S. Constitution declares in Amendment XII that

“if no person have such majority, then from the persons having the highest numbers not exceeding three on the list of those voted for as President, the House of Representatives shall choose immediately, by ballot, the President. But in choosing the President, the votes shall be taken by states, the representation from each state having one vote”.

The Vice President is chosen by the Senate
“if no person have a majority, then from the two highest numbers on the list, the Senate shall choose the Vice-President;”

Below are fourteen maps, each with a 269-269 tie in the electoral college. These are permutations on changes to 12 states (OR, NV, AZ, NM, MN, WI, IA, OH, PA, WV, NH) and ME Congressional District 2. I’m certain that additional permutations are possible (for example, no scenarios below are shown with Kerry winning Missouri).

Tie Map 1 Tie Map 2
Tie Map 3 Tie Map 4
Tie Map 5 Tie Map 6
Tie Map 7 Tie Map 8
Tie Map 9 Tie Map 10
Tie Map 11 Tie Map 12
Tie Map 13 Tie Map 14

As a note, Congress could easily eliminate no electoral majority scenarios in future elections (due to a tie) by changing the size of the House of Representatives to an even number (edited due to error pointed out by Dennis).

Voting simultaneously for Nader and Kerry?

There is an interesting article in the New York Times (May 5, 2004) by Bruce Ackerman, a professor of law and political science at Yale, titled 2-for-1 Voting that proposes cross-endorsement of Presidential Electors between Nader and Gore.

As quoted from the Article:

“electors will be named by each state’s political parties. But Ralph Nader is running as an independent. When he petitions to get on the ballot in each state, he must name his own slate of electors. While he is free to nominate a distinctive slate of names, he can also propose the very same names that appear on the Kerry slate.

    If he does, he will provide voters with a new degree of freedom. On Election Day, they will see a line on the ballot designating Ralph Nader’s electors. But if voters choose the Nader line, they won’t be wasting their ballot on a candidate with little chance of winning. Since Mr. Nader’s slate would be the same as Mr. Kerry’s, his voters would be providing additional support for the electors selected by the Democrats. If the Nader-Kerry total is a majority in any state, the victorious electors would be free to vote for Mr. Kerry.”

Given the lack of Instant-runoff voting (IRV) for Presidential Electors, this is a possible alternative to the “wasted vote” problem. Currently, the process of fusion is practiced regularly in New York as well as occasionally in several other states. Fusion is the process of combining votes from several party lines into a single total. In New York, for example, 2000 featured Bush on the Republican and Conservative ballot lines while Gore was on the Democratic, Liberal, and Working Families ballot lines (Buchanan was also listed under two parties – Right to Life and Buchanan Reform). The choosing of Presidential Electors is based on the fusion (or combining) of the votes for each ballot line of the candidates. However, this is generally practiced for only the same candidate. The legal possibility of different candidates cross-endorsing the same set of electors and allowing fusion is likely an issue to be determined by individual state laws.

As an historical note, cross-endorsement was actually somewhat common in the late 19th century. For example, in 1892, a set of electors in North Dakota were cross-endorsed by both the Populist and Democratic parties. In a very interesting result, two of these fusion electors and one Republican Elector received the highest number of total votes, creating the only occasion of a state casting electoral votes for three different parties (one Democratic, one Republican, and one Populist).

I’m interested in hearing any legal analysis of this possibility – especially in “battleground” states.

Super Tuesday

Today is Super Tuesday – states that collectively account for about 33% of the US population vote today in primaries (and one caucus), including the large states of New York, Ohio, and California. Massachusetts also votes today. The election here is a modified open primary (meaning that “Unenrolled” voters, i.e. Independents, can vote in either party’s primary), I headed off to the polls to cast my ballot – this time in the democratic presidential primary (I will resubmit my voter registration card to change my registration back to unenrolled shortly). Not many people at the polls when I voted – and it was a spectacular day, weatherwise, here in eastern Massachusetts (in the 60s Farenheit! whoa!). I cast my ballot uneventfully into the optical scanning machine and headed out. Most the major Democratic candidates were still on the ballot, including those that dropped out early, such as Gephardt. Waiting now to see how the results turn out.