LONG BEACH, Calif. -- In local elections, the field of candidates can be expansive.

Sometimes, especially in a winner-take-all special election, there is barely a clear victor.

In the 1st District City Council special election in April, Robert Garcia won with 40.7 percent of the vote, beating six other candidates but falling short of a majority.

Sixth District Councilman Dee Andrews received even less of a mandate when he won office in a May 2007 special election. He beat six other candidates with 27.25 percent of the vote.

In regular elections, the primary often has similarly low numbers, which is why the election is followed by a runoff between the top two candidates. In Long Beach's April 2006 primary election, no one received a majority of the vote in the races that had more than two candidates.

City Clerk Larry Herrera says a new election system, called instant runoff voting (IRV) or ranked choice voting, could ensure majority wins in a single election and would save the cash-strapped city millions of dollars.

However, critics say IRV is wrought with problems that destabilize the very heart of democracy. Supporters say IRV improves democracy, makes more votes count and lets voters avoid having to choose "the lesser of two evils."

"It's something that deserves, I believe, some consideration," Herrera told the City Council's Budget Oversight Committee this summer. "It would be up to the policy makers to Advertisement Quantcast determine its benefit, but the money savings are there."

On Tuesday, the policy makers - that is, the City Council - will try to determine whether IRV has enough benefits to warrant being placed on a ballot for voters to consider a city charter amendment. The council has a special meeting at 3:30 p.m. in City Hall, 333 W. Ocean Blvd.

How IRV works

Under IRV, voters don't just vote for one candidate, but can vote for up to three.

Voters rank their candidates in order of preference, and these lists are used to create the "instant runoff" within a single election. There is no primary, no runoff election, only a single election to determine the winner.

Once the initial election votes are tallied, the candidate with the least number of votes is eliminated. The voters who ranked that candidate as their first choice then have their second choice count toward a different candidate.

If that second-choice candidate is eliminated, then the voters' third choice applies. The process continues with every candidate, eliminating them one-by-one and tallying new votes for those who remain, until finally one candidate has more than 50 percent of the votes.

To implement IRV in Long Beach, the council must vote by Jan. 15 to place a city charter amendment on the April 2010 ballot. Then, more than 50 percent of voters would need to approve the new system for local elections.

IRV would first be used in the 2012 races for council, school board and other local seats, unless a special election occurs before then.

Long Beach wouldn't be the first city in the United States, or even in California, to use IRV, although it would still be in the minority.

According to election reform advocacy group FairVote, eight cities and Arkansas use IRV, while several others will soon implement the system. San Francisco first used IRV in 2004, and Oakland and Berkeley have adopted it and will use it for the first time in elections over the next two years.

However, IRV has had its controversy and its criticisms.

Pros and cons

In a lengthy report to the council for Tuesday's meeting, Herrera outlines some of the advantages and disadvantages of IRV.

IRV would save money in a city that faces ongoing budget deficits. Herrera estimates a cost savings of $3.7 million over eight years after initially spending $300,000 to $330,000 to upgrade ballot-tabulating computer software and spending $150,000 to $180,000 on voter education.

Beyond the financial benefits, IRV proponents say it gives voters more choices by encouraging more candidates to run, fosters better debate and less mudslinging, and increases turnout. Proponents say voters prefer IRV, largely because it reduces voter fatigue by cutting the number of elections.

In Long Beach, that last advantage certainly would have been welcome over the past three years.

From the beginning of 2007 to the end of this year, voters in some parts of Long Beach will have gone to the polls 12 times.

This is in part because of special elections to fill vacant seats, but also because of having primaries and runoffs for local, state and national elections.

However, as Herrera notes, critics say IRV can be too complicated for some voters to understand, that the first-choice candidate of a majority of voters won't always win, and that it is too complex and time-consuming to implement and count votes. They even say that, in some cases, picking a candidate as the first choice could actually hurt his or her chances of winning.

Gautam Dutta of the New America Foundation, an election reform group that has been pushing for the council to consider IRV, refutes these criticisms.

He said that IRV encourages more candidates to challenge incumbents or non-mainstream candidates because of the importance of second and third choices. IRV forces candidates to cross political lines and build coalitions, he said.

"Under the current system, you call up a voter and see if they want to vote for you. If they don't, you write them off," Dutta said. "That no longer works under IRV."

Dutta also says IRV ensures that the candidate with the most support will win an election, but critics note some theoretical cases that seem to contradict that.

Kathy Dopp, a mathematician for the National Election Data Archive, wrote an analysis that concludes instant runoff voting causes more problems than it fixes.

She notes cases in which a candidate that in a simple majority should lose to all but one other candidate could actually win an IRV election because of the elimination process. Also, she says IRV is "unstable" when three or more strong candidates are in a race, tending to give a minor fringe candidate the win over one preferred by a majority of voters.

Terry Reilly, a San Jose resident who has become a vocal opponent of IRV and has sent council members information to persuade them away from considering the system, notes other problematic examples. In one case, ranking a candidate first on a ballot can actually have the unintended consequence of bumping that candidate from the running.

In some cases, candidates benefit more from being ranked No. 2 by some of their voters, which Reilly says goes against Democratic principles.

"When a voter goes into the voting booth and pushes that name as No. 1, it should not hurt that candidate," Reilly said. "And this way, it hurts that candidate."

Dutta discounts these issues.

"Basically, they have theoretical validity, but no real life validity," Dutta said. "No one has ever come up with an election where any of these things that they're talking about have come up in real life."

Another flaw that Dopp notes is that not all ballots are treated equally under IRV because some voters' selections aren't included in the final count to determine a winner.

For example, in San Francisco's Board of Supervisors general election in 2004, 22 candidates were on the ballot.

Of 35,109 eligible votes cast, the winner ultimately received only 13,211 votes, or 37 percent. However, the 13,211 votes were 50.5 percent of the remaining votes following 19 rounds of eliminations, redistributing second- and third-choice votes and recounting.

Voters who chose three candidates who were eliminated early essentially had their votes taken out of the race by the end of the process.

Interestingly, the candidate who initially had a plurality - the most votes - in that election also won a majority with IRV.

Dutta said this lack of a true majority is no different than in traditional elections, when voters who support minor candidates are forced to choose between the two winners in the runoff, even if they may not want either one. The runoff system also discourages people from voting for the candidate they actually want to win if they think that person doesn't have a chance, he said.

With IRV, a voter could vote for, say, a far-left environmentalist who has little political clout, but then choose for a second choice a more mainstream Democrat that is also acceptable to the voter.

The runoff system also hurts voter turnout because people usually have to go to the polls twice. In Long Beach's major local elections of 2002, 2004 and 2006, voter turnout ranged from 14 percent to 22 percent in the primary, but from 20 percent to 28 percent in the runoff.

Still, a look at San Francisco's voter turnout over several decades reveals no significant increase in overall voter turnout after the implementation of IRV.

Dutta admits IRV has some problems, but he says it is still better than holding runoff elections.

"No system is absolutely perfect, but the question is which system has the least number of drawbacks," Dutta said.

What it would mean

Let's go back to those large fields of candidates in past Long Beach elections.

There's no way to know what voters' second or third choices would be in any given race, but let's consider the 5th District council primary race of April 2006.

Then-incumbent Jackie Kell finished in second place in a write-in campaign against now-Councilwoman Gerrie Schipske, who later won the runoff election. In the primary, Schipske and Kell beat six other candidates with 33.9 percent and 23.8 percent of the votes, respectively.

Under IRV, instead of going to a runoff election, the second choices, and in some cases the third choices, of the supporters of the remaining candidates would be factored in. What might that have meant?

Perhaps many of the losing candidates' voters also liked Schipske, and she still would have won her seat with a majority. Or, maybe more voters would have put down Kell as their second choice, giving her a victory.

Even odder, candidates who had a decent showing, such as third-place Ed Barwick, and who received strong second-choice support from the other losing candidates' voters could actually end up winning.

In this example, Barwick could move to second place or even first by riding the second-choice votes. This wouldn't give him a majority but would force Kell and maybe Schipske down the list and one of them out of the race, depending on their own second-choice support.

If Barwick had enough second-choice support from the new third-place finisher's voters, then he would win.

That is one of the ironies of IRV that some opponents criticize. The candidate who has the majority of first-choice votes doesn't always win.

However, it's exactly what supporters say is one of the benefits of IRV - it forces candidates to build coalitions and cross political lines because those second- and third-choice votes could make or break the election. It also means a candidate who is acceptable to the majority, even if not everyone's first choice, could win.

Council members have expressed mixed opinions about IRV. During a meeting last month, some members said they didn't even want to have a study session on IRV at all.

Councilwoman Tonia Reyes Uranga has expressed concerns about IRV potentially disenfranchising some voters. Councilman Patrick O'Donnell has been against the idea since the Budget Oversight Committee, which he is part of along with council members Gary DeLong and Suja Lowenthal, recommended a council discussion on it this summer.

"I know the issue that we're talking about is money, a very important issue, but I don't know that we alter democracy because it costs more to hold an election in the traditional American style," O'Donnell said at one of the committee's meetings.

Along with DeLong and Lowenthal, Councilman Robert Garcia has been a strong advocate of considering IRV.

"I think we need to do all of our research, get all of our questions answered," Garcia said last week. "From what I've read so far, I like instant runoff voting."

Paul.eakins@presstelegram.com, 562-499-1278