Tuesday, December 27, 2005

Election reform by Krist Novoselic

A Washington State System?

This is a guest post by Krist Novoselic, a board member of the Center for Voting and Democracy and a Washington state resident.

There is a battle in Federal Court regarding our primary elections. The struggle is between two principles. But Washington State need not choose between free association and more choices with our elections.

Ranked Choice Voting is a versatile electoral system with many benefits. It’s an idea worthy of the Pierce County Charter Review Commission’s consideration.

Here is an overview of where things are and where they can go.

Cajun Primary

I-872, the Top-Two Primary passed by a 59.84% majority. Voters wanted a primary with more choices that was like the late Blanket Primary. The restrictive Montana Primary, tired out by voters mere weeks before the general election ballot question, was unpopular and voters spoke accordingly in November.

It’s likely the Top Two Primary will be restored – but with ballots featuring candidates lacking any party designation.

The US Supreme Court set the stage for non-partisan primary elections when it struck down California’s Blanket primary (1). In that ruling, Justice Scalia, writing for the majority on the court, said a non-partisan Blanket Primary, “ … has all the characteristics of the partisan blanket primary, save the constitutionally crucial one: Primary voters are not choosing a party’s nominee.”

The cosmetic remedy of erasing party designation off the ballot comes at a high cost to voters. The valuable information voters get by the little D, R or whatever affiliation next to a candidates’ name is important. A non-partisan ballot deprives voters of this informative tool.

Big Sky Liabilities

Let’s face it; while the Montana Primary does protect a political party’s rights, it does little to speak to the valid frustrations of most Washington voters.

Many voters perceive the Montana system as a benefit only to the major political parties at the expense of taxpayers.

Most elections are uncontested or uncompetitive in our state – and primary elections are even more so. The Montana system only aggravates this sad state of affairs. Few or no choices on a partisan ballot can make voting virtually meaningless.

Don’t be fooled by thinking Washington voters will become comfortable with the Montana primary over time. I-872’s formidable near 60% poll showing will embolden its resurrection in non-partisan form.

Ranked Choice Voting – A Friendly Alternative

We need not adhere to the false dichotomy between more choices and free association currently slogging through public discourse regarding our elections.

On October 25th 2005, a group of Washington State citizens, myself included, came together with the Center for Voting and Democracy, to file a friend-of-the-court brief (2) in the appeal before the US 9th Circuit Court, that seeks to overturn I-872, the top-two primary.

All of the parties in the case seek to block the Court from considering this valuable information, ostensibly because of a supposed sense of urgency or that, greater voter choice, vote dispersal problems and freedom of association are somehow irrelevant to this lawsuit.

But most Washington voters are not in any hurry to participate in primary elections as evidenced this past September when 63% shrugged off primaries.

By filing this brief, we recognize the deeper challenges regarding our state’s elections. Too many citizens hold our democratic structures in low esteem. Most races are uncontested or uncompetitive. And primary elections only tend to make the situation worse.

Our brief seeks to draw the Court’s attention to a small modification of the top-two system that would;
• alleviate the concerns of all parties involved,
• remove the constitutional problems created by I-872,
• ultimately build a quality democratic system that would invite increased participation

Modifying I-872 with a ranked choice ballot would still narrow the field to two candidates who, regardless of party affiliation, then advance to a general election.

The main difference is that each voter ranks the candidates for office in his or her order of preference, from the first choice down through as many candidates as the voter chooses to rank, or as many as the rules permit. It’s as easy as listing 1 – 2 – 3. Votes are counted in a series of rounds. Each round eliminates the candidate with the fewest votes and redistributes ballots to those voters’ next choices. This process of redistribution and elimination is repeated in subsequent rounds and the top two ranked candidates advance to the general election.

Ranked choice voting accommodates the elimination of the self-identification provision that doomed I-872. Parties could be allowed to nominate their own candidates. However, in order to re-create the wider choice benefits of I-872, parties would be allowed to nominate multiple candidates for office. With ranked choice voting, parties could be assured of their votes converging on their most popular candidates while avoiding dispersion problems. Parties would hence have great incentive to nominate a diverse field of qualified candidates from within their party in order to bring a wider swath of their supporters to the polls.

Partisan voters could rank their candidates faithfully, while at the same time, independent voters who do not wish to vote along any party line may now select candidates from a larger mixed pool thus preserving the wider choices of I-872.


A Ranked Choice Blanket Primary promises many more benefits. In San Francisco, where Ranked Choice Voting is used, negative campaigning was greatly reduced. In fact, rival candidates actually endorsed each other, vying for their opponent’s supporters second choice.

Ranked Choice Voting fosters positive participation because voters express themselves without feeling like they’re throwing their vote away or wasting their vote – a tragic reality in the lexicon of American democracy.

Washington voters could get better value out of our primary elections. Most races are uncontested thus poorly attended. Why not utilize this smart and efficient system in otherwise mostly redundant elections?

As we mention in our brief, this is a system that can be built on. Down the road, we can choose to fold the primary into one efficient ranked choice general election. This will shorten the election season and save tax dollars. Pierce County can lead the state with this important electoral innovation.

Ultimately, we can consider the Proportional Voting version of ranked ballots that would;
• make our county and legislative races much more competitive,
• give voter’s real choices
• increase the voice in Olympia of candidates and voters around our state who are politically handicapped because of redistricting.


Legal precedents show Ranked Choice systems as fully consistent with the 14th Amendment of the US Constitution (3), and also complying with the Voting Rights Act (4).

No provision of the Washington State Constitution poses any barrier to the ranking of candidates in a primary. Ranked Choice Voting does not interfere with voters’ guaranteed right to free and equal election as it does not frustrate the voters’ will and allows each voter an equal opportunity to rank their choices. In 1908, the Supreme Court of Washington upheld a Ranked Choice system very similar to the one proposed here for our state primary (5).

Foes of ranked ballots in San Francisco tried hard to slow its implementation after it was passed by voters. But they ultimately failed and ranked choice voting stands unchallenged because of its solid case law. Once put into place, voters were wildly enthusiastic about more real choices. All the benefits did come to fruition.


Pierce County needs to consider ranked ballots. This innovation speaks to many of the problems currently weighing on our democracy.

When looking at all the uncompetitive or uncontested races, lack of participation and voter cynicism the question begs to be asked – is this the best democracy can get?

The Charter Commission and Pierce County voters must ask themselves this question.

Krist Novoselic

1. California Democratic Party v. Jones 530 U.S. 567 (2000)
2. Amici Curiae Brief Of Fairvote – The Center For Voting And Democracy et al. See http://fixour.us/editorials/11.1.05.html
3. Wesberry v. Sanders, 376 U.S. 1, 18 (1964). Moore v. Election Comm’rs of Cambridge, 35 N.E.2d 222, 238 (Mass. 1941)
4. Thornburg v. Gingles, 478 U.S. 30, 47 (1986).
5. State v. Nichols, 50 Wash. 508, 97 P. 728 (1908). See http://wahdems.org/rcv.html


At 6:53 PM, Blogger Nathan said...

I've posted a rather lengthy response on my blog which you can read here.

I first heard about Instant Runoff Voting on NPR shortly after it was implemented in San Francisco. It intrigued me, and I was inclined to support it. The more I learned about it, however, the more concerned I became.

At this point I can't support its adoption, either in Pierce County or in the state as a whole.

At 4:27 AM, Anonymous Anonymous said...

"Ranked Voting" (IRV) will only be used to increase voter choice.

That means IRV can be used in primaries to choose each party's candidate,
and in the general election to choose between the already chosen candidates of each party.

"Proportional Representation" is DOA, give it up.

At 4:43 PM, Anonymous Anonymous said...

Instant Runoff Voting, the system you describe in your post, unfortunately suffers from some major flaws.

One is that it's next to impossible to tally up IRV ballots and process them centrally; all the counting has to be done at local polling places, with election HQ calling back after each elimination to get a new count, or else all the ballots have to be collected at one location. In contrast, under the system we use today, you can simply add up the number of votes for each candidate at local polling places, send those to central election HQ, and figure out the winner there.

Another is that IRV will choose a winner in many situations that goes against voters' preferences - it gives much more value to their #1 choice than to their preferences in general. When most people prefer candidate A over candidate B, you don't expect candidate B to lose to candidate A. If you're going to have ranked choice ballots, you ought to use the information provided by voters to choose a compromise candidate - one who satisfies the most preferences.

If you're looking for election reform, I have two suggestions: (1) approval voting. This is a simple modification of our current voting system, where the only difference is that you can vote for more than one candidate. Whichever candidate is "approved" by the most voters is the winner. (2) Condorcet voting. This uses the same ranked ballots as IRV, but the counting method is greatly improved, and it will select a compromise candidate as the winner when possible.

At 1:10 AM, Anonymous Anonymous said...

Jesse is simply lieing.
about IRV theoretically choosing a less popular candidate. That is realistically impossible, but anti-reformers keep repeating it to try to confuse newbies.

The plurality system doesn't work for voters, so it doesn't matter how easy it is to count the votes. Modern elections are counted by computer, so his 1st paragraph is ridiculous fear-mongering too.

He sure fit a lot of bs into one message though. Get a life! So-called "approval" has not and will never be considered for real govt. elections except by fools with ADD trying to waste the world's time.

Please keep your bs to blogs about "approval" voting instead of wasting our time and lieing about real reform. That message should be deleted from the blog if you don't want to waste readers time.

At 9:55 AM, Blogger Nathan said...

Actually, Jesse is not wrong about the failure of IRV to pick compromise candidates. I've swiped the following example from the wikipedia entry on instant runoff voting

Imagine that candidates are located along a one dimensional ideological spectrum, and that the center of the spectrum is defined by the median voter. IRV does not reliably choose the option closest to the center of the spectrum. Thus it can be argued that IRV is less apt at choosing compromise candidates (than for example the Condorcet methods), and more likely to choose an ideologically polar candidate.

For example, this failure can occur in a 3-choice election where parties A and C are bitterly opposed, and party B is first choice for a minority but tolerable for a large majority. For a real-life example, consider the 17th-century Europe struggle over "government-enforced Catholicism" versus "government-enforced Protestantism, with "freedom of private worship" as the compromise B.

Imagine that votes are cast as follows:

38% of voters - 1. A, 2. B, 3. C
11% of voters - 1. B, 2. A, 3. C
13% of voters - 1. B, 2. C, 3. A
38% of voters - 1. C, 2. B, 3. A

In IRV, the compromise (choice B) is eliminated immediately. Choice C is elected, arguably giving severely lower total satisfaction amongst voters than choice B, who is preferred by a large majority to A, and who is also preferred by a large majority to C.

Of course our current system fares only a little better in that regard, but at least it doesn't introduce the temptation to use minor parties as preference farms.

In any case, I doubt those minor parties would fare much better under an IRV system than they do now. My analysis is that we'd reap no significant benefits from such a change, and would inherit an entirely new crop of problems.

By the way, I'm at least theoretically amenable to either of Jesse's alternatives, though some serious thought would need to go into tie breaking before I'd be entirely comfortable with it.

I think Jesse oversimplifies Condorcet voting a bit. To begin with, it's not a voting system per se, but a criteria for analysing election systems.

In a given election, a Condorcet winner exists if there is one candidate who would not lose in a head to head match-up with each of the other candidates. Such a candidate might not command a majority, or even a plurality, but he would be preferred by a majority to any other candidate.

The Condorcet winner in the example I cited above would be B, because matched up against either A or C it would win by 62% to 38%.

Not every election will have a Condorcet winner, which is where things get tricky and tie-breaking becomes important, but the mark of a Condorcet compatible voting system is that it always chooses a Condorcet winner when one exists.

I'm rather fond of the Schulze method myself, which is a Condorcet compatible system. It's somewhat similar to Jesse's approval voting.

IRV is not a Condorcet compatible system. It is much more likely to favor polarized candidates rather than compromise ones.

At 12:05 PM, Anonymous Anonymous said...

Soon, all of Washington will close polls and vote by mail. My county does. Mailed ballots, (a paper trail) are sent to the county auditor – a central location. This new reality accommodates Ranked Choice Voting (RCV) quite nicely.

People tend to dwell on the non-monotonic aspect of Ranked Choice Voting. But this is only an anomaly – which doesn’t merit disqualifying this promising electoral reform, nor the greater benefits it provides.

To be fair, let’s look at the status quo voting systems and their anomalies.

In 1996, during the Republican Blanket Primary, Ellen Craswell won with less than 20% of the vote. Mrs. Craswell was a conservative State Senator who wore religious beliefs on her blouse sleeve. Her views on social issues didn’t fit well with Washington Blue State sensibilities. As a result, Gary Locke practically waltzed to the general election. And as far as discussing important issues, all Locke had to do was say he wasn’t Craswell.

The Blanket Primary failed because the vote fractured in the large field of Republicans (Nine candidates, I believe). There were three well qualified moderate Republican’s who split most of the vote between them. Fringe candidate Craswell’s small plurality granted her victory in the primary – without even considering the votes of the more numerous mainstream voters. Craswell wasn’t the only loser in the general election, moderate Republican and independent voters also lost out because of the weak dynamic of the campaign.

Let’s hypothesize that the 1996 gubernatorial race was conducted under the Montana primary. It’s a good bet that the results would be very similar, (It’s hard to say how many Democrat crossover or mischief voters strategically voted for Craswell, hoping to doom her in the general).

And what if we put the 1996 election through the prism of either a partisan or non-partisan Top Two primary? The November election would have featured only King County Executive Gary Locke and Seattle Mayor Norm Rice.

Defenders of the status quo will tell us these situations happen infrequently. And they’re right. The non-monotonic issue with RCV is trotted around like it’s going to upset every election. But in reality it’s an anomaly potentially appearing no more frequently than the situations mentioned above. Even though none are perfect, RCV offers more benefits than either the Montana or Top-Two system.

I find some of RCV’s most persistent critics coming from the Condorcet community. I’ve quit an IRV newsgroup due to constant haranguing of some Condorcet supporters. They’ve even bums-rushed the Wikipedia IRV site. At a public hearing in Olympia last year, a staffer from the Secretary of State testified their “skepticism” about RCV. This staffer read a disparaging quote, notably from a pro-RCV or IRV website. We found this quote attributed to a Condorcet supporter. Personally, this was disappointing because I like Condorcet.

The Charter Commission should consider all voting systems. Contact State Representative Toby Nixon. He is very sharp on elections and a Condorcet advocate. Work with him on introducing a bill in Olympia.

Please don’t dwell on the monotonic issue only to wreck RCV. RCV has a great track record and is gaining momentum around the nation. Pierce County will benefit from a system like Ranked Choice Voting.

At 5:52 PM, Blogger Nathan said...


You're right that the status quo is not much better in terms of non-monoticity than IRV. I admitted as much in my previous comment. (You're also wise to be wary of wikipedia, by the way, though that doesn't address the validity of the example.)

My stand on the issue is this: unless you can show that IRV provides significant and tangible benefits over the status quo that are not offset by corresponding disadvantages, I will prefer the status quo.

You've shown that there are benefits, but my analysis leads me to believe that they are either not significantly better than the status quo or that there are corresponding disadvantages that offset them. Do we really want to spend the time and money necessary to implement IRV if it's only going to be a great leap sideways?

At 6:46 PM, Anonymous Anonymous said...

Nathan, you are way out there.
IRV passed by 84% in Takoma Park, the 4th landslide in a row.

Ranked Voting reform ended the Irish civil war, giving Ireland the strongest economy in Europe.

IRV was preferred more thqan 3-to-1 by SF voters according to the latest large-scale survey.

IRV has been supported unanimously by many councils this year. In Berkeley, the two incumbents who were voted out recently ( a VERY rare occurrence) were defeated largely because opposing IRV revealed them as dishonest, dumb, or simply out-of-touch. That includes a two-term mayor and CA longest serving election official.

By duissing IRV instead of spending a few hours to learn the facts of how ranked voting solves election problems, you are just wasting your time and looking less than intelligent.

(ftr, the main benefits of IRV are probably ending negative partisan campaigning and virtually ending the influence of campaign money in politics)

In the long run, it doesn't really matter what IRV opposers think- they are a tiny minority just wasting everyone else's time, and they always wise up or crawl back in their holes to be forgotten. It is a SIMPLE election modernization.

Sorry to hurt your ego, but facts are facts.

At 7:18 PM, Anonymous Anonymous said...

Krist is right about Condorcet (aka "IRV+").
Back in '03, one of our councilmembers was buffooned by the now defunct anti-IRV "ElectionMethods" website and he repeated a bunch of lies claiming IRV would pick the wrong winner.

Talk about eating his words. After the 72% landslide, the councilmember has gone to huge efforts to repaint himself as a pro-IRV expert.

In REAL elections, "Condor..."/ranked pairs (aka IRV+) and IRV would offer the same benefits to everyone and the SAME WINNER.

The problem with IRV+ is that it is practically impossible to count without computers, so it would be impossible to pass anytime soon.

At 7:22 PM, Anonymous Anonymous said...

IRV is nearly impossible to count without making repeated phone calls to every district, or shipping all the ballots to a central location. Ranked pairs can be tallied at local precincts and added up centrally, which means faster returns.

"IRV+" is a misleading way to describe the voting method I suggested, because it doesn't involve instant runoffs at all. Remember, IRV is an instant simulation of regular runoff elections, eliminating the least popular candidates until there's a clear winner. Ranked pair voting looks at *all* voter preferences simultaneously.

And you think it's "practically impossible" to count? Please. Counting ranked pair votes by hand just requires a little knowledge. It's not some black box like an electronic voting machine; anyone who's been to high school can learn to count ranked pair votes in a few minutes. Furthermore, you said yourself that "modern elections are counted by computer", so your point cancels itself out.

And of course, counting approval votes is as easy as counting the plurality votes we use today. You complain because approval voting hasn't been used in government elections, but perhaps if people like you were more interested in choosing a good election system and less interested in shouting down your opponents--you actually suggested my message should be *deleted* because you disagree with it--then maybe it would've been used by now.

At 4:31 PM, Anonymous Anonymous said...

Calling "Ranked Pairs" ANYTHING is the real farce.
Please take your stupid sci-fi ideas somewhere else. IRV is moving forward nationwide with the only stupid delays from idiots and liars claiming that there "lost-in-space" so-called voting system that has never been passed anywhere is somehow better, based on sci-fi hypothetical ballots that would never be cast in a real election.

Get a life,
and stop trying to interfere with real reform

Jesse is just another me-tooer trying to deceive the naive. Your idea has no following and no hope of success. You think you look smart, but even the newbies will figure out and realize that you are just a troublemaker trying to get attention by interfering and slandering ideas more respected than your bs.

At 5:28 PM, Anonymous Anonymous said...

There was a time when members of a political party could enforce discipline on elected officials of their pary and that was the strength of party identification. Parties fashioned platforms and expected elected officials to basically follow them or suffer withdrawal of party support. Voters had a general idea of how candidates would vote, based on the party platform. Elimination of parties through non-partisan elections and changes in how candidates campaign have tended to diminish reliance on the local party and, therefore, the responsiveness of elected officials to the persons electing them.


Post a Comment

<< Home