Election 2004 by the Numbers
Election 2004 by the Numbers
We've been sifting through the results of the November 2nd elections. They tell important stories - ones that in some cases have been overlooked or misinterpreted by many observers. I think you'll enjoy perusing through our findings below.
I also wanted to report on three landslide wins for instant runoff voting at the ballot this November. Instant runoff voting (www.fairvote.org) is rapidly growing in popularity as a means to elect majority winners when more than two candidates contest an executive / one-winner office.
* Proposal B on Ferndale, Michigan's ballot won by a lopsided 69%-31% margin. The proposal amends Ferndale's city charter to provide for election of the mayor and City Council through the use of IRV pending the availability and purchase of compatible software and approval of the equipment by the Ferndale Election Commission. A suburb of Detroit with some 10,000 voters that are relatively balanced between Democrats and Republicans, Ferndale had a very energetic, effective campaign led by Ferndale IRV: www.firv.org
* In Vermont, voters in Burlington overwhelmingly passed an advisory referendum on whether the city charter should be amended to use IRV for the election of the mayor. Under Burlington's current charter, a candidate for mayor can win with as little as 40% of the vote (meaning 60% might consider that candidate the worst choice), and if no candidate achieves that threshold, a separate runoff election is held. These provisions offer the worst of both worlds, creating the risk of a "spoiler" scenario and also the potential cost and lower turnout typical of a separate runoff. Some 66% of voters approved the ballot item, meaning that a formal charter amendment is likely to move forward in March.
* Voters in 16 western Massachusetts towns approved a non-binding motion in support of IRV, by a margin of 11,956 to 5,568. The question directed state representative Steve Kulik to vote in favor of legislation or a constitutional amendment to require IRV for elections to statewide office (such as Governor, Treasurer, Auditor and Secretary of the Commonwealth
The final good news on the instant runoff voting was San Francisco's first IRV election. Despite introducing the system to voters in the midst of a presidential year, the city reported a smooth transition. First-choice results were reported on election night. With absentee and provisional ballots being integrated into the totals, initial runs of the IRV program should take place on Friday -- in the future we expect quicker results, and cities and states that require all absentee votes to be in place by election night could run IRV tallies that evening. For a San Francisco Chronicle news article, see: http://fairvote.org/sf/sfchronicle110304.htm
Before turning to our "Election 2004 by the Numbers", I will make one point about the election process in this country. Many observers are suggesting that the election went smoothly. Although we applaud all the election officials, observers and alert voters who helped make our elections work better than in they could have been, we would politely disagree that having only 71% of our adult population registered to vote and forcing some voters to wait in lines that take more than 10 hours are signs of a well-operating electoral process.
More fundamentally, I believe we aren't hearing as much about problems in significant part because this year one state isn't holding the future of the presidency in an election requiring a recount. If Ohio had been 100,000 votes closer, we suspect we would be hearing hourly stories about controversial practices, the "chads" that are used on Ohio's many punchard machines, why there were so many provisional ballots, how overseas ballots were handled, double-voting and the like. We continue to have a patchwork of laws and practices that are an ongoing accident waiting to happen.
We are developing a series of recommendations for congressional action to protect our citizenship right to vote, starting with a right to vote in the Constitution and continuing through statutory changes such as universal registration to ensure clean and complete voter rolls, making Election Day a holiday to ensure both an adequate pool of pollworkers and increased access for voters, and uniform standards for voting equipment. We can -- and must -- do better, and we would be foolish to become complacent.
Onto our report on "Election 2004 By the Numbers." Our key findings include:
* The 2004 election was in fact a very status quo one, reflected by the near exact Electoral College mirror of 2004 to 2000 and the almost perfect stasis in U.S. House races. Even the Senate gains from Republicans fit into this pattern, with all Republican gains coming on ground that already was firmly Republican in 2000. Of course when Republicans control the White House and Congress, a status quo election is a victory for their party.
* The House of Representatives has reached a breathtaking level of non-competitiveness. More than 95% of seats were won by margins of more than 10% - a record. Only four incumbents outside of Texas didn't win by at least 4%, and only three were defeated. The House has changed partisan control only once since 1954 - and unless Republicans suffer major setbacks in the 2006 midterm election, it almost certainly won't change hands anytime soon. This lack of competition is partly due to redistricting, partly due to incumbent advantages, partly due to campaign finance - but primarily due to the fact of winner-take-all elections in single-member districts. We support full representation voting methods as the one indispensable part of any reform package seeking to provide real choices and fair representation to all voters.
* Our Monopoly Politics projections in US House races were extremely accurate on victory margins. Made without any attention /to campaign financing and candidate behavior and using a one-size-fits-all model, we projected landslide wins in 211 seats - and 210 those seats indeed were won by 20% landslide margins. Of the 13 seats we identified as most vulnerable with our model, fully 7 changed parties - among only 11 of 435 seats that changed overall. Only six seats changed hands in 403 seats outside of Texas.
Here is more detail on our findings for each level of election:
- George Bush certainly ran more strongly than in 2000, a year in which he received a half million fewer votes than Al Gore. This year President Bush's popular vote victory margin will likely be about 3.5 million votes - and was much larger in total numbers of votes received due to the rise in participation. His percentage of the vote rose consistently by 2-3% in most states, reflecting a general rise in the national tide of support -- although one that Democrats countered to some degree in such battlegrounds as Iowa, Ohio and Wisconsin.
- At the same time, 48 of 51 the Electoral College contests (in the 50 states and the District of Columbia) voted for or against Bush according to how they had voted for Bush in 2000. A shift of only 35,000 votes in Iowa and New Mexico (Bush's narrowest wins in 2004 and Gore's closest wins in 2000) and New Hampshire (Kerry's closest win in 2004 and one of Bush's two closest wins in 2000) would have resulted in all 51 contests going exactly as they had gone in 2000.
- If Bush's victory had been smaller - perhaps by one million votes instead of three - John Kerry likely would have won Ohio and thus the Electoral College and the presidency. That win would have meant two consecutive dysfunctional presidential elections where the popular vote winner did not win the presidency. This year's race easily could have gone to a 269-269 tie, after which the U.S. House would have picked the president, with one vote per state - a tie would have occurred if Kerry had won a total of 46,000 more votes in Iowa, Nevada and New Mexico (and perhaps a good deal less once all the provisional ballots are counted).
- For those dismayed by how the presidential campaigns so clearly focused all their energy and resources on the 16-18 states defined as battlegrounds, watch out. If anything, the number of battlegrounds likely will decline in 2008. If this year's national vote had been a 50-50 tie and the vote share had changed equally across the nation, only 5 states would likely have been decided by less than 4%, and only 15 states by less than 8%. Democratic states in fact are more solid than Republican ones in this scenario - a tie vote this year certainly would have elected John Kerry based on this year's results. Thus, don't expect more inclusive presidential campaigning in 2008 - and quite possibly an even smaller one, with all attention again paid on the two big truly swing states, Florida and Ohio.
- For Republicans to win all 50 states, their candidate likely would need to win more than 63% of the national vote. (Republicans can forget completely about winning in Washington, D.C., where Bush in 2004 did not crack double digits). A similar vote share for Democrats would likely win only 42 states; to win all 50 seats, their candidate likely would need to win more than 70% of the national vote. These sharp differences reflect how the nation's partisan polarization is very real. Exit polls suggest that George Bush won only 10% support from African-Americans (11% of all voters) and John Kerry won only 23% of evangelical Christians (22% of all voters).
* U.S. Senate
- Republicans had a net gain of four seats in the Senate, but there are important caveats about the mandate in that result. First, in U.S. Senate races Democratic candidates overall won approximately three million more votes than Republicans. Second, Republicans only gained seats in states that George Bush had carried in 2000 at the same time he lost the national popular vote -- Florida, Georgia, Louisiana, North Carolina, South Carolina and South Dakota. Third, five of their six seat gains were in open seats without incumbents, and each of the winning Republicans in these open seat races ran behind George Bush's winning total in the state.
- The sixth seat gain for Republicans was in South Dakota, where Tom Daschle was defeated by less than 5,000 votes (and where he and his opponent John Thune spent more than $30 million in an election where 390,000 votes were cast - more than $75 per vote). Daschle was the only Senate incumbent to lose; the Democrats' two gains were in open seats in Republican-leaning Colorado and Democrat-leaning Illinois.
*U.S. House of Representatives
- This House election was the least competitive in history. 416 out of 435 seats (95.6%) were won by non-competitive victory margins of at least 10%. 369 out of 435 seats (84.8%) were won by landslide margins of at least 20%. More than 99% of incumbents outside of Texas won, with only three (one Democrat and two Republicans) losing. (Four Democratic incumbents lost in Texas after being victimized by brutal gerrymandering, as detailed below, including two losing to Republican incumbents.) Only one victorious incumbent won by less than 4%. Note that these safe incumbents won in an election where the voter turnout was 50% higher than it had been in 2002 -- but the new voters broke along very similar partisan lines, based largely on the partisan nature of most districts.
- George Bush's coattails were very limited. Outside of Texas (see below for more on the impact of that state's 2003 gerrymander), Republicans picked up only two seats in the U.S. House and lost four. Republicans defeated only one Democratic incumbent (by 1,365 votes in a district that George Bush likely carried by more than 45,000 votes) and gained only one open seat, winning by 31,000 in a district that Bush likely carried by 70,000 votes. All but two of the remaining Democratic incumbents won by margins of at least 10% -- and those by the relatively comfortable margins of 7% and 9%. Only five Democrats, including those defeating incumbents and winning open seats, won by less than 7%, and only one won by less than 4%. Republican targets among incumbents in 2006 are quite limited.
- Open seats went heavily to the party that had already been holding that seat - 29 of 33, with one of those seat changes in a much-changed district in Texas. Of those 33 seats, 30 went to the candidate of the party whose presidential candidate had carried the district in 2000.
- Tom Delay's Texas gerrymander was immensely successful for Republicans. Democrats lost no seats in the 2002 elections after the 2002 redistricting, resulting in a delegation that was 17-15 Democratic. Today, in the wake of this week's elections in the 2003 plan, the delegation is 21-11 Republican, a shift of six seats. Just as conceived by the plan's architects, white congressional Democrats were decimated, reduced from 10 in 2003 to three. Of these three, one (Edwards) won by just 4% in his heavily Republican district, and the other two represent Latino-majority districts. By 2012, it is quite possible that no white Democrat will represent Texas in Congress.
- In November 2002, within days of the election, we issued our "Monopoly Politics" projections for November 2004 House races, for which we needed to know absolutely nothing about campaign financing, the quality of challengers and incumbent voting records and behavior. The only changes we have made since then were factoring in the 33 open seats and the 32 seats changed in the Texas redistricting plan. Once our one-size-fits-all formula was adjusted with that information, we projected 211 landslide winners of at least 20% -- and 210 indeed did win by landslide. We projected another 107 comfortable wins of at least 10% - and 105 indeed did win. We projected another 33 winners - and 32 won. Yes, despite missing only four projected margins out of 351, we did have two of our projected winners (Phil Crane in Illinois and the open seat in Colorado's CD-3) defeated - making three errors out of more than 1,600 projected winners in the five House elections starting in 1996.
- Washington state voters adopted (even as California voters rejected) a version of Louisiana's "top two" system. This year's elections were the latest example of the quirks of this system. In Louisiana, all candidates run on the November ballot. If no candidate reaches 50%, the two top vote-getters face off in December. (In Washington, the first round will take place in September, with the top two always facing off in November.) There will be two hotly contested runoffs this December in competitive seats in Louisiana, both with one Democrat and one Republican. In CD-3 all Republican candidates won a total of 59% of the vote and all Democrats won a total of 41%. But the third-place Republican candidate finished less than 2,100 behind the second-place Democrat, with another 10,300 votes going to a Republican who lagged behind - the December contest thus easily could have been between two Republicans. In Washington State, we suspect third party candidates will almost never now be able to contest the November election, and key races will regularly lack a candidate from one of the major parties.
* Women, racial minorities and third parties:
- Women increased from holding 60 U.S. House seats to 64 seats, just shy of 15% of the House, A woman candidate has a solid chance of winning one of Louisiana's two runoff elections in December. Women maintained their 14% of U.S. Senate seats and will drop from nine gubernatorial seats to seven or eight depending on whether Christine Gregoire wins her undecided Washington State election.
- After gaining no U.S. House seats in 2002 after redistricting, African- Americans gained three new House seats in Texas, Missouri and Wisconsin. Asian Americans gained a new seat in Louisiana, and Latinos a new seat in Colorado. After six years without an African-American or Latino in the U.S. Senate, African American Barack Obama won in Illinois and Latinos Ken Salazar and Mel Martinez won in Colorado and Florida. White men and women now hold 49 of 50 gubernatorial seats and 95 of 100 Senate seats.
- Third parties had a sharply reduced impact in the presidential election, with the total third party under 1%. Third parties also had limited impact on congressional races, with only two victorious Senate and House candidates apparently held below 50%. Third parties increased their number of seats in state legislatures, but primarily in Vermont, where the Progressives now hold six seats.
* Governors and state legislatures
- Gubernatorial elections continue to be the single-most competitive level of election in the United States. Fully half of all states have had a governor from a new party in the past four years. Four of the 10 governor's races that have been decided changes parties - the 11th race in Washington is too close to call.
-According to the National Conference of State Legislators, Democrats gained 76 state legislative seats around the nation and picked up more legislative chambers than their Republican counterparts. As a reflection of a 50-50 nation, Democrats lead by just 12 seats out of a total of 7,382 seats. We reported this fall that only 61% of state legislative seats were even contested by both major parties See: http://www.ncsl.org/programs/press/2004/pr041103a.htm (NCSL news release) http://fairvote.org/reports/uncontestedraces.htm (uncontested races)
* Voter turnout
- According to Curtis Gans and the Committee on the Study of the American Electorate (CSAE), voter turnout (not counting those who made mistakes in their votes for president) will likely end up being more than 120 million adults, which is 59.6% of eligible citizens - the highest since 1968, when 61.9% of turnout and up from 2000 (54.3%), 1996 (51.5%) and 1992 (58.1%). Voter turnout rose in all but one state (Arizona). We will post CSAE's report on Friday, November 5.
- Turnout in the presidential battleground states increased by 6.3%. Turnout in the other states increased by only 3.8%. Turnout in noncompetitive New York rose by only 0.8%, while in hotly contested Florida and Ohio it rose by more than 8%.
At least three states voted at higher rates than the part of the United States that in 2000 and most other recent years has had the highest turnout in the nation: Puerto Rico. Puerto Rico is not allowed to vote for president despite its people being American citizens, but it again had a hotly contested race for governor, resulting in turnout of 70.5% of eligible voters. According to CSAE, this year's turnout was only higher in Minnesota (76.5), Wisconsin (73.7%) and New Hampshire (71.6%) and may ultimately be iin Oregon and Maine. Helping to explain its high turnout, Puerto Rico makes voting a holiday and has legislative elections that allow small parties to win seats through full representation. Minnesota, Maine, Wisconsin and New Hampshire all have election day registration. Oregon has vote-by-mail.
Thanks for reading, and stay tuned.