Democracy's most important problems and fixing them with Range Voting

Warren D. Smith

Free Press readers already know about blatant election manipulation tactics in Ohio and how J. Kenneth Blackwell simultaneously served as Bush-Cheney Ohio campaign chair and Ohio's highest election official, then supervised his own 2006 governor run.

Although Strickland defeated Blackwell by getting 60.5% of the votes, the Columbus Dispatch's poll, ending the Friday before the Tuesday election, predicted Strickland would get 67±1.2%. The Dispatch noted that this "7-point [discrepancy] broke a string of five straight gubernatorial elections in which [our] poll exactly matched the victor's share of the vote." (The theoretical chance of a statistical discrepancy this large is below one in 3 million.)

In comparison, a transfer of only 1.3% of the Ohio 2004 vote from Bush to Kerry would have given Kerry the presidency, and indeed CNN's exit poll at 12:21AM on election night predicted Kerry defeating Bush in Ohio by a 4.2% margin.  The magnitude of these election manipulations (if such there were) was 7% (or less) of the vote.

But there are other democracy-destroying problems of larger importance which have received far less attention, perhaps because they are too obvious.

Gerrymandering:  In Ohio statewide in 2006, between 53 and 61% of voters
voted Democratic depending on the type of race (US House, Senator, or Governor), but only 39% of the congressmen they elected were Democratic! Conveniently, all 4 landslide races (vote count over 65%) were won by Democrats, while all 9 House races closer than 60-40 were won by Republicans. That's called "pack and crack" and it's a far larger distortion than squelching a mere 7% of the voters.

Lack of voter choice and huge inherent distortions:   In the Ohio 2004 race, the Free Press's own Bob Fitrakis also ran for governor on an election reform platform (Green Party), collecting 1.0% of the vote, as did Libertarian Bill Pierce (1.8%).  Of the four candidates, only Fitrakis was pro-choice on abortion. But surely more than 1% of Ohioans care about abortion rights and election reform? (10% of Ohioans have had abortions.)

In the 2004 presidential race, all candidates besides Bush & Kerry put together got less than 1% of the vote, despite the fact that Bush and Kerry both voted for the Iraq war, subsidy-laden Farm Bill, and PATRIOT act, and both supported the drug war, WTO, and NAFTA, whereas all the alternative candidates (and roughly half the US population) were against them. In Spain, the Socialists won control (with Socialist S.Royal touted as the best shot at the French presidency) while the top USA socialist got 0.003% of the vote. Yes, French and Spaniards socialize more than than Americans, but maybe twice as much, not 20,000 times!

These numbers represent enormous distortions of democracy which completely shut out important viewpoints. Polls indicated that throughout June 2004 to now, 54-59% of Americans thought the Iraq war was a mistake – but neither major party candidate (the only ones who could possibly win) opposed the war. So voting for an anti-war candidate, such as Ralph Nader, was generally considered a wasted vote that would help elect the worst major party candidate. The same thing happened in the 1968 presidential election/Vietnam war.

What is going on here? One answer is massive 2-party domination caused by our voting system. Our voting system makes it strategically foolish to vote for any candidate besides the two majors – even if by far your favorite – because that "wastes your vote." Indeed "vote splitting" can cause the candidate least-liked by a voter majority, to win. Presumably for those reasons, about 90% of voters who told pollsters they thought Nader was the best candidate in 2000, actually voted for somebody else. That distortion is a huge obstacle for candidates like Nader. Consequently strong candidates usually refuse to run third party. Over time, all this causes third parties to die out. Political scientists call that "Duverger's law." Voters call it "getting the minimum possible amount of choice" – indeed it's usually worse than that since US elections are 98% predictable over a year ahead (far worse than any comparable democracy – in India, it's below 50%).

Ballot access laws: Not only have third parties been marginalized by the voting system, they often can't even get on the ballot thanks to a state-by-state crazy quilt of non-uniform and often amazingly-restrictive ballot access laws. Joe Lieberman's successful 2006 Independent run to re-win his senate seat in CT would have been illegal in 46 of the 50 states. In the approximately 400 US House races in Georgia since 1943, zero third-party candidates ever succeeded in getting on the ballot, thanks to huge petitioning requirements omitted for the two major-party candidates.

How to fix it:
1. Congress can and should enact uniform ballot-access standards. Because in all US history since the age of preprinted ballots began (about 1890) up to 2005, there have only been two cases where a ballot for some statewide office (or US presidential race) has ever had more than 10 candidates, provided at least 2500 signatures were required to get on ballot (and even in those two cases, there were less than a dozen candidates), we recommend that be the sole requirement in any race for statewide office.
2. The "shortest splitline algorithm" is a simple way to draw districts. It eliminates all bias because the district shapes are uniquely determined by the population distribution and shape of the state.
3. "Range Voting" is: each voter scores each candidate on an 0-to-9 scale (it is permitted to leave some unscored) and the highest average score wins. It's used by Olympic judges. Interestingly, it's also been used for millions of years by honeybees.
• Switching to range voting is easy: all voting machines can handle it with no modifications and no reprogramming. (Even easier is "Approval voting" which is the "0 or 1 only" form of range voting where voters just approve or disapprove each candidate. That's actually simpler than the current voting system because it just discards the "no overvote" rule.)
• It gets rid of the "spoiler" and "vote splitting" effects: voting your favorite top is never strategically unwise and if you do it you can still express your opinions of all the other candidates in your vote.
• Because of that, it should gradually allow third-parties to enter the picture.
• Your vote expresses the most you can, not the least you can.

To begin to see how much distortion our democracy is suffering by not using Range Voting, consider these facts:
• Nader was a nonfactor in 2004 with 133 times fewer votes than Bush, but our exit poll found he would only have been a factor of about 1.6 below Bush's total with Range Voting.
• A study by Brams and Fishburn found that Anderson would probably have come in second (behind Reagan but ahead of Carter) in 1980, if Approval Voting had been used.

These huge distortions far outweigh a piddly 7%.

A wrong "fix":
"Instant runoff voting" (IRV), while it may initially sound attractive to the uninformed, is inferior to range voting.
• Won't work on many present-day voting machines and can't be counted in precincts.
• Voting your favorite top is often strategically unwise.
• Spoiler effect still is present (just in a more subtle form) and still leads to massive 2-party domination in every country in which IRV has been tried, e.g. Australia. Meanwhile the closely related genuine-runoff system usually has not led to 2-party domination. For this reason, it is foolish for, e.g, the Green Party, to support IRV – and especially foolish to support replacing genuine runoffs with IRV.
• Greatly increases the risk of election-tie nightmares.
• Experimentally increases ballot spoilage and uncounted votes. More complicated than range voting.
• Tends to favor "extremists" and prevent "centrists" from winning (say computer simulations).

Warren D. Smith is a math PhD and co-founder of . Click here to learn about the fact checking behind this article.


Return to main page