Monday, March 16, 2009

Why don't they ask people if they would prefer IRV when it costs more than traditional elections and runoffs?

I am getting so sick of people claiming everyone prefers IRV to traditional elections and runoffs, citing surveys as proof IRV is the greatest thing since sliced white bread!

I just attended a meeting of the Cary Town Council and there were some pretty awful assumptions made about IRV - that it's always cheaper than holding traditional majority elections with runoffs if needed.

From the

The respondents were next asked their support for using the Instant Runoff Voting Method using a 9-point scale from not supportive at all (1) to very supportive (9). The respondents were also informed the use of the method would save Cary taxpayers approximately $28,000 by not having to hold a physical runoff election.
While supporters are claiming that Cary taxpayers saved $28,000 in the 2007 election by using IRV, they cannot honestly make that claim because they didn't keep track of all expenses. IRV was done under the table and off the books. Labor and services were donated by non-profit groups (including pro-bono services from so far un-named PR firm and free websites provided by the main group pimping IRV) and none of it was kept track of.

But even public information officers seem to buy into the very simplistic argument that one election is cheaper than two. Check out the fiscal impact statement at

Fiscal Impact:

(1) Below are the FY 2010 budget estimates based on each type of election, with and without IRV (all costs include one-stop early, contract costs, etc.). These estimates include the IRV estimates that the Town of Cary would absorb to cover voter education. These estimates assume the maximum number of elections possible with each method of election. Also, these figures include the cost of Cary contracting with Wake and Chatham Counties to conduct its municipal elections, since Cary is in both counties. These numbers reflect the highest possible amount of charges we would incur; actual expenses would more than likely be less as the county pays a portion of the cost of elections when other county issues are on the ballot.

Partisan Primary and Election without IRV (potential of three elections):


Partisan Primary and Election with IRV (two elections and IRV education costs):


Non-Partisan Plurality without IRV (one election):


Non-Partisan Plurality with IRV (one election plus IRV education costs):


Non-Partisan Election and Runoff without IRV (two elections):


Non-Partisan Election and Runoff with IRV (one election plus IRV education costs):


Non-Partisan Primary and Election without IRV (two elections):


Non-Partisan Primary and Election with IRV (one election plus IRV education costs):


(2) Fiscal impacts for future budget years might potentially be less if the Town were to continue using IRV, especially once IRV is well established and no longer requires public education.
That last statement is incredible, because there is ample evidence from the San Francisco Board of Elections that shows they have been spending $1.87 per registered voter on IRV elections since 2004, and in 2008 a civil grand jury informed them that they aren't doing a good job educating voters. Even after 4 years, voters not only require continuing education but they haven't been getting enough even after spending $1.87 per registered voter! See for more information.

What about the people who continue to move into Cary - will they not need IRV education? What about people who become citizens or turn 18 - will they not need voter education? It's overly simplistic to

It has not been proven that IRV will save any jurisdiction any money. In fact, financial impact studies by other governing bodies such as in VT and MD and even Washington state have shown that IRV could easily cost more money and be much more complex to administer and be less transparent and open.

There are also no real goals here – one of which is whether or not IRV is easier or harder for voters to understand how to cast their votes? Is another goal of the IRV pilot program to compare how easy or difficult it is to count or otherwise administer IRV elections vs. traditional elections and runoffs?

Several jurisdictions have performed studies to see how much it will cost to implement IRV. The Vermont Secretary of State performed a study. VT is a small homogenous state with highly educated and active voters and no central elections commission or state-wide standards like we have in North Carolina. Even though VT claimed they could do IRV on the cheap with voter education costs around 25 cents per registered voter, their study did find that IRV would be more expensive.

Maryland is a state with a central elections board and of a similar size (slightly smaller) and makeup. The MD Legislature performed three fiscal studies on IRV – in 2001, 2006 and 2008 – and estimated as best they could the costs per registered voter for initial implementation ($3.54) and ongoing voter education ($0.48) on top of the current costs per registered voter for election administration. These costs were not complete because there is certified software or equipment that can tabulate IRV at this time. None of the IRV bills passed out of committee, presumably because it appeared that IRV would cost more than the cost of rarely-needed runoff elections..
What would the combination of the Maryland costs per registered voter for implementation and for voter education be if they were used here in NC?

Initial cost of $3.08 per registered voter multiplied times 6,283,277 registered voters. 48 cents per registered voter in ongoing costs every year thereafter.

Initial cost of $3.08 per registered voter multiplied times 6,283,277 registered voters. 48 cents per registered voter in ongoing voter education costs every even year thereafter, and 24 cents per registered voter in odd-year elections.

The Maryland cost estimates factor in elections every two years, whereas in North Carolina we have local elections in odd years, and elections for county officers in both odd and even years. Therefore IRV voter education would need to be done each and every year. With 5.8 million voters over the summer and 6 million voters as of 11/04/08, annual voter education costs alone would equal the cost of the rarely needed state-wide runoff elections. Add to that nearly $20 million in implementation costs (not including certified software and hardware that does not exist) and IRV would never be cheaper than holding second elections even at the state level. So unless you blindly believe that one election (no matter how complicated) is cheaper than two, the push to IRV can’t really be about costs unless you chose to ignore this evidence.

So why is it that no one is asking questions about supporting IRV over traditional elections and runoffs even if IRV costs more than those elections?

Saturday, March 14, 2009

Report on the 3/12/09 Cary Town Council Public Hearing on IRV

On 3/12/09, the Cary Town Council held the second meeting where the 2009 IRV pilot was addressed.

There was about a two to one ratio of pro-IRV speakers to anti-IRV speakers. Andy Silver, Perry Woods, Don Hyatt and myself spoke out against IRV. Don Hyatt, a Republican from Cary, used similar words used by John Hollingsworth (former president of the Wake County Progressive Democrats):

"The IRV math doesn't make sense!"

Here is Andy Silver's report:
Chris, Perry, I, and a couple other people spoke against having a second IRV pilot in Perry at the Town Council meeting tonight, and about twice the number spoke for. Then Erv Portman made a motion for IRV, and spoke for it, but Don Frantz gave a devastating account of the problems of counting the ballots in 2007, followed by comments against IRV by 3 or 4 other members. Finally, Erv withdrew his motion and another motion to use plurality to decide elections was passed, meaning only that allowing plurality elections will be discussed again, and I think there will be an open forum on it April 15. This does not rule out IRV, which can be discussed again at the same time (really complicated, in parliamentary terms). Several members wanted to get clearer information from the BOE (not sure whether Wake or state) on possibility of counting the second and third rounds of votes by optical scan - also whether the sorting before those counts could be done by machine.

Some remarks were made about disappointment with the BOEs for not providing analysis and conclusions about the 2007 pilot election - otherwise what is the point of having a pilot?

I thought that the discussion by council members was amazing - could provide an excellent civics lesson on IRV and its pitfalls. I look forward to their posting the video of the proceedings on the town website.

It was clear from talking to some town council members before the meeting that they were impressed by our side's information. The pro IRV people commented mostly on the reported ease of voting, and how Cary voters liked it according to the exit poll. They seemed to feel that our side had the most reasoned arguments.

So at the meeting I wasn't surprised at the pro-IRV arguments I heard. The same old stuff was trotted out - IRV ensures majority winners (it did not do so in Cary), it costs less (not if you keep track of all the costs and don't do it under the table and off the books), and people like it (but do they really understand it?), etc. The Cary Town info officer showed some information was cheaper, but I suspect that was not including other costs that other legislative groups have factored in.

In fact the whole issue that Cary is still claiming that IRV elections are cheaper disturbs me, given that there is overwhelming evidence from many sources (San Francisco, VT, MD, and now Pierce County WA) that IRV elections are more expensive than traditional general elections and runoffs once all the costs are considered. But Cary continues to claim that a single IRV election is always cheaper than a traditional runoff election.

From the

The respondents were next asked their support for using the Instant Runoff Voting Method using a 9-point scale from not supportive at all (1) to very supportive (9). The respondents were also informed the use of the method would save Cary taxpayers approximately $28,000 by not having to hold a physical runoff election.

No one really kept track of all the expenses in the 2007 pilot - including keeping track of the value of all the volunteer work done by FairVote and DemocracyNC that they actually have to keep track of in the 2009 pilot guidelines. So since the pilot was done so haphazardly in 2007, maybe they shouldn't claim that IRV saved money over traditional elections?

They should ask the question a couple of different ways:

  • If using IRV in a single election cost more than holding a traditional general election plus runoff, would you favor using it in Cary?
  • If using IRV meant we had to pay for the cost of voter education and an exit poll, and that cost more than holding a general election and runoff election, would you favor using it in Cary?

Another thing that was amazing about this survey was that the results showed there was a difference in how people answered the questions based on age, race, gender, income, education level, and where they lived (apartment, house, condo, etc.) - something that the original exit poll did not reveal. Could that be because Cary actually hired a professional polling organization that didn't have a dog in the hunt?

Respondents who were registered voters were subsequently asked their understanding of the Instant Runoff Voting Method (Table 63). A 9-point scale was used ranging from do not understand at all (1) to understand very well (9). The results indicate there was a level of misunderstanding among the respondents. The mean was 5.83 with 58.6% on the “understand” side (above 5) of the scale and 30.6% on the “not understand” side (Figure 19). This includes 22.0% who indicated they do not understand at all. Overall this indicates a degree of misunderstanding among the respondents.

And here is more interesting information:

There were several subgroups indicating considerably lower levels of understanding including apartment dwellers (3.65), 0-1 year Cary residents (4.10), and those with no internet access (4.62). Also exhibiting a degree of misunderstanding were $30,001-$50,000 income level (4.96), 18-25 age group (5.00), 2-5 year Cary residents (5.25), townhouse/condo dwellers (5.33), 6-10 year Cary residents (5.52), $70,001-$100,000 income level (5.52), and those without a college degree (5.55).

So clearly IRV is not the wonderful thing unaffected by socio-economic boundaries as claimed by FairVote!

But after all that - it was obvious that everything presented by the Town of Cary and IRV supporters dealt with the front-end of elections - the voting. Although a good chunk of people didn't understand IRV, and 25% of voters in the FairVote survey didn't know they were supposed to rank their choices, none of the info dealt with the back-side (or ass-end) of elections - the election administration and vote-counting.

And on this point, having two sitting council members (Frantz and Robison) who attended the 2007 IRV tabulation along with Perry Woods and myself was a major factor in why three other council members (Adcock, Robinson and Smith) had such problems with it.

First off, they didn't like the fact that Cary was once again being asked to be a guinea pig for IRV - why not let someone else do it? There was no information on what was proved in the 2007 pilot - which was something I had been asking to be done even before the election. So if they had no idea what the 2007 pilot proved, why then do another pilot in Cary? Good points for sure!

Then Frantz and Robison stated quite cleary something that I have been stating all along - the IRV tabulation procedures are lousy! As simple as some claimed they were, the Wake BOE couldn't follow them - and this was the council members saying it on the record, not just me. So if they couldn't follow them in 2007 - how could they be sure the same procedures would be used and followed in 2009? The big problems were with the hand-sorting by board memners and volunteers with no overhead projectors, along with the math and calculator errors - and of course the secret non-public recount that no one (not even the candidates) were invited to!

Add to that some last-minute plan to somehow use M-100s scanners to count the hand-sorted votes that had no details to it, and the council members were not very impressed. They wanted more details - and I suspect those details either don't exist yet, or if they do - using the M-100s in that way might not be legal. The Council members weren't impressed by any plan that still used problematic hand-sorting. So they decided to ask for more information from the Wake BOE and deal with it at another meeting in two weeks.

Other than some things that I heard from one particular council member that I couldn't believe he actually said and believed himself, that was the end of IRV at this particular meeting. Then the Council voted 6 to 1 to consider switching to the lower cost plurality election method - which will be cheaper still than using IRV with any method. Especially since IRV only delivers plurality victories based on the total number of votes cast in the 1st round - which if IRV supporters are going to be intellectually honest about it - is all that should count if they want to stick to their claims that IRV is one single election. You can't go changing the denominators of an election to manufacture a majority win if you want to still claim it's a single election.

But this did remind me of the 2007 Raleigh City Council meeting where I went up against Bartlett, Gilbert, Poucher and Hall on IRV. The Raleigh City Council decided to wait two weeks to gather more info, then came back and couldn't even gather support for a motion to use IRV - it died in chambers. When municipal leaders really have the time to study IRV, and they have access to all the information - it goes down in flames.

That is what happened in 2007. And last year when the City of Wilmington was considering a special law to allow them to use IRV (before there was a pilot extension bill), they backed out of it because they found out about the back-side problems as well as the majority failure issues and costliness. It was also because someone wrote the bill to require the use of IRV, not give them the option.

It's just like it says on my 2nd plate at the Flying Saucer: "VAX IS USSR" (anagram)

Chris Telesca

Wednesday, March 4, 2009

2nd IRV election in Burlington VT does not result in a majority winner!

Got a posting on another news group from Anthony Lorenzo, Florida IRV advocate who appears to post damn near everything issued by FairVote.

FairVote is the nation's leading pusher for IRV propaganda. Here is a link to the FairVote release in the Huffington Post:

Burlington’s Has Second Highly Successful Instant Runoff Voting Election

This year, with so many viable choices, no candidate won an outright majority of more than 50% of first choices – indeed, plurality winner Kurt Wright won only a third of the initial vote. The election went to an instant runoff tally. In the instant runoff, the candidates with the fewest votes were dropped, including independent Dan Smith and Democrat Andy Montroll, and the field was narrowed to two finalists. In the final instant runoff round, every ballot counted as a single vote for whichever of the two finalists, Progressive mayor Bob Kiss or Republican Kurt Wright, was ranked higher on each ballot. By 8:25 p.m. the IRV tally was completed and Kiss had been re-elected, defeating Wright in the final round by 51.5% to 48.5%.
Notice that the comments for that posting are closed - apparently Rob Richie doesn't want to have people post opposing viewpoints about IRV - or that his claims are false.

The press release claims that the second IRV election in Burlington VT was a success and that incumbent mayor Bob Kiss had a majority win over Republican Kurt Wright 51.5% to 48.5%.

Is that true? Notice that the total number of votes is not listed anywhere in the article. I had to go to the Burlington Free Press to get the numbers for the first round and the final round:


Initial Round:

* Kurt Wright (R) 2952 - 32%
* Bob Kiss (P) 2585 - 28%
* Andy Montroll (D) 2065 - 23%
* Dan Smith (I) 1307 - 14%

So that is a total of 8909 votes

But in Burlington you need a majority to win and after two rounds of IRV, it was Kiss on top with 51.5 percent to Wright's 48.5.

Final Vote:

* Bob Kiss (P) 4313 - 51.5%
* Kurt Wright (R) 4061 - 48.5%

Bob Kiss had 4313 - or 48.41% of the original 8909, not 51.5%.
Kurt Wright had 4061 - or 45.58% of the original 8909, not 48.5%.

That is because the total number of votes for these two candidates in this round is 8374 - or 535 less than the original 8909 cast in the first round. That is why an IRV win is not a true majority win in all but one or two cases because you never really get a true majority of the first round votes cast.

Here is what happened - according to the Free Press:

According to an unofficial tally, Wright led after the first round of vote counting by 252 votes. After the second-choice votes of the fourth and fifth place finishers — independent Dan Smith and James Simpson of the Green Party — were redistributed, Wright still led by 213.

Wright lost when the 2,554 second-choice votes of third-place finisher Democrat Andy Montroll’s went to Kiss by a 2-1 margin.
How does Andy Montroll get 2,554 second-choice votes when he only got 2065 in the first round?

“I’ve heard so many people say this instant runoff system is flawed,” said John Pijanowski, a Wright supporter, after consoling Wright with a hug. “I hope he challenges this all the way to the highest court.”
Yes - I agree - it seems very flawed. And it does appear to advantage the incumbents because in rounds beyond the first, you might not know anything about the other challengers, so you might vote for the incumbent for no reason other than they already got the job. That is an advantage of incumbency.

On the other hand, had there been a separate runoff election and some campaigning between the top-two candidates, who knows who would have come out on top? While people who preferred Montrol as their first pick chose Kiss 2 to 1 over Wright in their 2nd or 3rd choices, we don't know the actual number of votes, or if they would have really picked Kiss over Wright if the race was only between them. That is precisely why RRO prefers traditional runoff elections OVER preferential balloting like RCV or IRV.

Even some colleges that vote on computers that don't count any of their votes the way we do in the real world admit that you have to take the total number of votes cast in the first round of an IRV race and use that as the threshold - you can't do what Burlington, Cary and so many others have done and have a gradually reducing threshold for your "preferential majority". Even the students at NCSU recognize that you might not have a true majority after all the ballots are exhausted - so you would need to have a second or runoff election. As we should have done in Cary and they should have done in Burlington.

Calling something a majority win when it is not a true majority is called many things - false advertising and fraud are two things that come to mind.