Wednesday, May 12, 2010

Latest IRV con in NC!

As I worked the polling places on May 4, I wondered how many races were going to require a runoff. I also wondered how long it would take before some folks would be pushing Instant Runoff Voting again.

Sure enough, one of my precinct voters told me that he heard DemocracyNC's Bob Hall talking about IRV on WUNC radio. I will find the link soon.

But sure enough, the first of what I assume will be many letters to the editor pushing IRV in very simplistic terms came out today - from Adam Sotak of DemocracyNC:

As your May 6 article noted, a high-cost, low-turnout statewide runoff is on the horizon. I remember one of these runoffs a few years back at which I was only the fifth person in my precinct to vote - and that was at 6:30 p.m.! There's got to be a better way. In fact, there is.

Instant runoff voting provides a cost-effective and simple solution. Voters go to the polls and rank their choices for an office: 1, 2, 3. (Nobody has to rank more candidates than he wants to.) In the first round of counting, only the first choices are tallied. A candidate who gets the prescribed threshold of first-choice votes (in this case 40 percent) is declared the winner.

If a virtual runoff is needed, all candidates except the top two vote-getters are eliminated, just as in the current system. If your first choice is in the runoff, your vote stays with that candidate. If your first choice was cut, your vote goes to the runoff candidate you ranked best. The candidate with the most votes wins.

Some cities in North Carolina have already successfully used IRV. It's time to expand this idea for statewide races.
Mr. Sotak is wrong when he states that IRV provides a cost-effective and simple solution to runoff elections. And now there is a new phrase to describe IRV - "Virtual runoff". That's a good idea, because you aren't having a real runoff. Nor does IRV give you a real majority.

IRV advocates would have you believe that having one election is always cheaper than having a traditional election and a runoff. That's not quite true. If you wanted to do IRV in NC, you would need entirely new voting systems because our current machines won't handle IRV without some serious jury rigging that makes our elections less transparent and verifiable. And our own State Board of Elections stated that IRV was too risky to use in the 2008 primary elections because they couldn't be made to comply with state and federal election regulations. So what's changed now, other than wanting to take short-cuts?

Voter would need voter education for the complex and sometimes confusing IRV method each and every year. That's not cheap. Other states have done fiscal studies of IRV and they have found that IRV would be a more costly voting method than having runoff elections. The Maryland legislature studied IRV twice and both times found it to be more expensive. If you took their cost and applied them to NC, it would cost us $18 million to implement and $4 million each and every year for voter education. All that just to save $3 million this year - and you'd never end up saving money with IRV!.

With more than 3 candidates in an IRV race, there's a chance that a voter wouldn't vote for a candidate who made it to the final count. So that voter's votes wouldn't help any of their candidates. And in an IRV race, the top finisher in the first round has a greater than 90% chance to win any subsequent round. So all IRV does is delay the inevitable. In a traditional runoff election, the second-place finisher in the first round goes on to beat the first-place winner 33% of the time. So Cal Cunningham would be more likely to lose with IRV than in a traditional runoff election, where voters would have a real chance to make another choice, based on any number of factors: endorsements from opponents in the primary, new information that comes out, more debates - all things that voters DON'T know about with IRV. IRV is hardly as democratic as a real runoff - even Roberts Rules of Order favors traditional runoff elections over IRV - or what they call "Preferential voting".

And it turns out that in over 90% of IRV elections that are settled beyond the first round, the "winner" doesn't even get as many votes as they would need to win the first round. That happened in Cary in 2007, where a winner was declared with only 1401 out of 3022 votes. Hardly the 1512 votes they would have needed in the first round. And there were many problems with IRV that one single time that we used it to count those 2nd and 3rd column votes!

Furthermore, IRV is only allowed in a limited number of cities per year as part of an election pilot. Only 2 cities used IRV in 2007 - Cary and Hendersonville. Cary Town leaders didn't like the confusion caused by the problems counting the 3022 IRV ballots, and didn't want to be lab rats in 2009. Hendersonville voters ranked their choices on IRV ballots in 2007 and 2009, but voters lucked out - winners were determined in the first round, sparing voters, candidates and election officials the fun of counting ranked votes from DRE touchscreen voting machines and then transferring them over to a Microsoft Excel spreadsheet for the final tally - a method that is not only not very transparent, but most likely untested for accuracy and is definitely uncertified.

There would be no way to count these votes in a statewide election without hauling all the ballots to Raleigh to have them counted. And even if there was, in a big primary election like we had in 2008, it would have taken so long to count the ranked votes that the results wouldn't have been tallied until AFTER the results of the traditional runoff election was known, So much for IRV being instant.

And many other places that got talked into using IRV are dumping it, some after only one try:

  • Aspen, CO (where there is now a criminal investigation as to whether some election laws were broken in order to do IRV in Aspen, including the illegal certification of the election by a private company that didn't have the power to certify an election);
  • Burlington, VT where more people voted to dump IRV than voted to pass it a few years before;
  • Pierce County, WA - where 63% of voters didn't like IRV and the costs for IRV were HALF of the county election budget.

I talked to other poll greeters for Republican candidates about the possibility of runoffs in their races, and told them to watch out for people pushing IRV. When I described how IRV works, and told them about my experiences observing the 2007 Cary IRV election and vote counting, they shook their heads in wonder - and heard more than one person say that this sounded like a con game. IRV doesn't work and it has many more problems than our current methods.

Brad Friedman - the well-respected blogger and election expert - wrote that IRV stands for "Instant Runoff Virus" I call it "Instant Runoff Voodoo"!

It's ENRON vote counting, and we don't need it here in NC.

No comments: