I support this. There's really no discussion to be had; it's the right thing to do.

by andrew on Feb 1, 2011 11:58 am • link • report

I agree with this proposal, but Cheh is beholden and protective of the privileged place that the DCDSC has. And so are most of her colleagues. They will cast this as an attack on the party, as if the rights of the party are more important than the rights of voters. Just look how apoplectic they got at the idea that people would change their registration in the weeks before the primary.

People rarely give up power and perogative willingly. You have to rip it out of their hands.

by TM on Feb 1, 2011 12:07 pm • link • report

IRV elects candidates with as little as 24% support.

Just like the upcoming special election with 21 candidates, San Francisco's recent IRV election for the District 10 Supervisor. Malia Cohen won with 4,321 votes when 18,508 votes cast ballots in that election. Over 75% of the ballots did not list her name.

http://www.docstoc.com/docs/70605705/Breakdown-of-the-D10-RCV-race-in-San-Francisco

by Jackson on Feb 1, 2011 12:56 pm • link • report

That's a lot more than if you had 21 candidates in a standard election, where the winner could have as little as 5% of the vote.

IRV isn't perfect. Mathematically, it's been proven that no voting system can be perfect. For example, on IRV, one person could be the #2 choice of everyone, the #1 choice of nobody, and get eliminated in the first round.

However, its flaws are much less severe than the flaws of the standard system, and most importantly, it is one system that makes it possible to switch from a primary-general system to a single election without introducing huge problems.

by David Alpert on Feb 1, 2011 1:01 pm • link • report

This is a somewhat tangential issue, but I also would suggest that DC should move to a system like Philadelphia has, where it is written into law that out of the seven at-large seats, no more than five can be from the majority party. Since Philly, like DC, is an overwhelmingly Democratic city, what this effectively means is that 2 of the 7 at-large seats are "reserved" for Republicans (or Independents, though in practice it's always been R's who win the two seats).

It's a blunt, if effective, way, of ensuring that no party can ever control 100% of the Council.

It may not ever happen, but it's still a good idea.

by Marc on Feb 1, 2011 1:04 pm • link • report

I'm beginning to see that Council Member Cheh is useless and should be ousted when her term is up for reelection.

by Rich on Feb 1, 2011 1:06 pm • link • report

The point is, IRV elects people that the majority of the population prefers to most of the other candidates, while opening the field to a broader group of people.

I would have liked to see the recent Ward 1 council member race results with IRV.

by RS on Feb 1, 2011 1:13 pm • link • report

@Marc We actually have that rule in DC. It's written into the Home Rule Act. Two seats on the 13-member council (someone correct me if I have that wrong) are reserved for minority parties. That includes "independents" like the perennial Democractic candidate who changes to (I) for strategic reasons, like Michael Brown, whose father was *chairman* of the DNC.

I don't blame people like Brown. This is just a stupid system. Whether we have all D's or a mix is irrelevant. Parties are mostly irrelevant in cities like DC.

by Ward 1 Guy on Feb 1, 2011 1:38 pm • link • report

+1 to IRV and just one trip to the ballot box. Let sanity prevail!

BTW, without a closed party primary you might see many registrations switch from (D) to Independent or (R).

by Ward 1 Guy on Feb 1, 2011 1:42 pm • link • report

I like it, but it's a system that requires a bit more voter involvement than now. Or maybe way more. It requires each voter to rank every candidate (or not, in which case their vote could become irrelevant). It's one thing to figure out whom you like the most, but do we really expect voters to be able to choose between their 20th and 21st preferred candidate? Keep in mind, that choice could be determining the election.

by ah on Feb 1, 2011 1:47 pm • link • report

You say in the opening paragraph that the election would be nonpartisan (meaning there would be no political parties). If this were the case, it would have the interesting side effect that a Federal worker could be elected to the DC council or as Mayor and not violate the Hatch Act.

by Steven Yates on Feb 1, 2011 1:48 pm • link • report

Let me emphasize that this post makes so much freaking sense that the only thing wrong with it is its length. We need a brief version to sway people who don't realize what a no-brainer this is.

by Ward 1 Guy on Feb 1, 2011 1:51 pm • link • report

I'm a fan of IRV, but, in all reality, the most important thing to do is move to a non-partisan election. If IRV is a roadblock, then scrap it. Everybody on one ballot (even if there's a run-off requirement) is a far, far better way of doing things.

by springroadintoaction on Feb 1, 2011 1:54 pm • link • report

ah: Someone who just ranks 2 candidates under IRV is still participating more than someone can under the current system.

Right now, you rank only 1. If your person isn't the top vote getter, it really doesn't matter whether you voted for the 2nd highest or the 21st highest.

Under IRV, if you rank just 2, then if either of those people ends up winning, your vote counted. If it comes down at the end to one of them versus one of the others, the fact that you voted for one of them helps that candidate.

If the top vote getters are all people you didn't rank at all, then under IRV you're not participating, but under standard voting you wouldn't either if the top vote getters are all not the person you voted for.

by David Alpert on Feb 1, 2011 2:00 pm • link • report

@Steven Yates: I've thought about that, but it still doesn't work. Federal workers can run for non partisan office, but they still can't fundraise. It's pretty tough to win a council seat without raising any money (or ever participating in an independent fund raising campaign).

by TM on Feb 1, 2011 2:31 pm • link • report

Approval voting is so much easier and better than IRV. If we were going to switch (and it strikes me as incredibly unlikely, but let's fantasize anyway), then I'd much prefer to see the nice simple Approval system. Ranking 20 politicians is a ridiculous way to vote, and IRV can have really strange unintended results.

by jcm on Feb 1, 2011 2:36 pm • link • report

I agree that the important point is that everyone goes on the same ballot regardless of party. City elections in California are like this (San Francisco's recently implemented IRV aside) and I believe this is true elsewhere as well. If someone gets more than 50 percent in the first round s/he wins, otherwise the top two candidates go to a runoff. You can make the "first round" the primary and the runnoff the Nov. election, or you can have the first round in November and then a special runnoff election later if needed.

I'm actually in favor of IRV but it's so offputting to so many people that I think a nonpartisan two-round system might be the thing to push for.

by jfruh on Feb 1, 2011 2:40 pm • link • report

No need for something as unnecessarily complicated as IRV.

Just make it open primaries; problem solved!

Of course, don't expect the Dem Party-beholden members of the City Council to ever propose such a thing.

by Fritz on Feb 1, 2011 3:40 pm • link • report

If not IRV, you could have it slightly modified. If no one gets 50% +1 of the vote, you need a run-off. If not, you avoid the entire, costly primary process and thereby save everyone money. Why should the city have to foot the bill for a primary?

by SJE on Feb 1, 2011 3:44 pm • link • report

How will states like Louisiana, which hold a runoff a few weeks after the November election, manage under the new rules?

I wonder whether the hurdle is too high if you try to institute IRV and scrap the two party system. Maybe pick either (a) a runoff of the top two vote getters from the general election or (b) IRV in the Democratic primary.

by JimT on Feb 1, 2011 3:49 pm • link • report

An alternative system would be "Approval Voting," where you get to cast a vote for as many candidates as you like.

There are cases (Board of Ed jumps out) where this sounds like it wouldn't be a bad idea. If you have feelings that "I wouldn't mind any of these 3, but don't want any of those 3 on the board," approval voting seems like a nice option, and likely jives with how most voters feel about those positions.

It also lets you do things like checking every box except for the ones belonging to Vincent Orange and Marion Barry.

IMO, Mayor & Council seats should be done with IRV; At-Large council seats and BoE should be done with Approval voting. Also, make all regular council-runners eligible for the at-large seat, and disqualify them if they win their ward's seat.

This would be the fairest and most straightforward system, if not a bit complicated.

by andrew on Feb 1, 2011 4:09 pm • link • report

IRV:

* Does NOT guarantee a majority; someone already pointed out SF district 10, but the Oakland mayoral race saw the same thing.

* Does NOT guarantee the winner is "preferred over all other candidates". (The actual guarantee is much weaker; that the winner will be preferred over at least one other candidate.)

* Does NOT allow you to simply vote your true preferences. With 3 (or more) strong candidates, the same "spoiler" problem is still possible, and the logical strategy is to, as is common now, vote for the lesser of two evils.

New Hampshire is examining a bill for approval voting, and there is talk of the same in Colorado. DC should do the same.

by Dale Sheldon-Hess on Feb 1, 2011 4:17 pm • link • report

What exactly is the strategy for getting IRV implemented? The city's political establishment is overwhelmingly Democratic - because that's how the voters like it. So Democrats are obviously the people who you need to support this. And I really mean the median DC Democratic voter - African American, older, a city native, and living in Wards 1, 4, 5, 7 or 8. It's their power you're advocating diluting. So what's your argument to them about why they should agree to this? And I doubt that adequately accounting for the preferences of independents and Republicans is an argument that will win them over. Basically, why should we go to system that would have re-elected Fenty, when that median Democratic voter didn't want to re-elect Fenty?

I don't think the arguments I see here are wrong, but voters (and the politicians who want their votes) are much more likely to act out of self-interest than fairness. So how do you convince those currently holding the balance of power, and who aren't in any danger of losing it, that they should support this?

by Peter on Feb 1, 2011 10:13 pm • link • report

Approval Voting is simpler than IRV and massively superior according to every objective metric.

www.electology.org/approval-voting-vs-irv

by Clay Shentrup on Feb 1, 2011 11:32 pm • link • report

Approval Voting is simpler than IRV and massively superior according to every objective metric.

This is just totally not true. By far the most common case is where there are only two candidates with any chance, and some spoilers. In that very common scenario in approval voting, you have to be strategic and identify who the real competition is, to vote effectively. With IRV you don't have to be strategic at all, you get the optimal result by just listing your true preferences.

IRV is much much much better in real elections.

by David desJardins on Feb 2, 2011 3:48 am • link • report

Every time we talk about IRV some people pop up who probably have a news alert for IRV and start bashing the system in favor of approval voting. It reminds me of whenever we talk about Personal Rapid Transit (PRT) and suddenly the pro-PRT and anti-PRT folks all swarm over to the blog.

If I wanted a lot of comments, I should post an article that says, "which is better, IRV or PRT" and let the fur fly!

by David Alpert on Feb 2, 2011 9:02 am • link • report

@ David desJardins That's not the way approval voting works. Approval voting allows you to vote for all candidates you find acceptable. So if the election is, say, Bush, Gore, and Nader, one could vote for both Nader and Gore without helping Bush or hurting Gore or Nader. No strategy necessary.

@ David Alpert Don't forget to mention the metro bag searches in your mega-comment post. Oh, and maybe a little gentrification, too. :)

by jcm on Feb 2, 2011 9:41 am • link • report

@jcm: That's not the way approval voting works. Approval voting allows you to vote for all candidates you find acceptable.

Yes, I know exactly how approval voting works.

Let's say the main candidates are Alice, Bob, and Charlie. You prefer Alice to Bob, and Bob to Charlie. So, you have to decide whether to cast your vote only for Alice, or for both Alice and Bob.

If Alice and Bob are the main candidates, and Charlie is going to be a runner-up regardless, then you need to vote for Alice alone. If you vote for both Alice and Bob, then your vote would be wasted and irrelevant to the outcome.

On the other hand, if Bob and Charlie are the main candidates, and Alice is going to be a runner-up regardless, then you need to vote for both Alice and Bob. If you vote for Alice alone, then your vote would be wasted and irrelevant to the outcome.

This is the main problem with approval voting. You need to be strategic and know not only whom you prefer, but know who has a chance of winning and who doesn't, and use that in your vote decision. It's asking a lot more of the voters.

by David desJardins on Feb 2, 2011 11:55 am • link • report

IRV is not the solution to DC's problems! Don't voters deserve to know which candidate follows the platform of their preferred party? And don't those same voters deserve to know which of their party's candidates deserve their support after winning a primary election? By going to a non-partisan primary, you just hide that information and require more money be spent on getting a message out that could be done cheaper through the party structure.

And take it from someone who has studied IRV and seen two IRV elections turn all those wonderful promises about IRV upside down! I live in NC, where we've done a few IRV elections: two pilots in 2007 with one IRV tabulation; one pilot with no IRV tabulation; and the grand finale to IRV in NC - a statewide IRV election for non-partisan judicial office!

The first IRV tabulation failed to deliver a majority winner (1401 out of 3022 ballots is NOT a 50% plus one vote majority win). The procedures were written in advance and planned for by the observers, but they were not followed. There were mistakes made by the counters that were caught by the observers, but not all the mistakes were caught - there was a secret recount where some missing votes were discovered. But that was on paper ballots - at least in the other IRV race in NC (in Hendersonville - the only IRV race in 2009) where they vote on touchscreen machines, which were going to be counted on uncertified excel spreadsheets in violation of state and federal laws.

But we had a mandate to use IRV for judicial elections under certain circumstances, which occurred in 2010. Now we had no certified software to handle the tabulation, nor had our General Assembly voted to appropriate any money to figure out how to administer IRV elections. So they jury-rigged it and bent our election laws to pull it off. Only the fact that most of the candidates took public money meaning you won't be able to any extra raise the money to challenge the results. And there is much to challenge - the counting methods differed from county to county, and even the non-federal testers admitted that there were other variables which could have affected the count.

NC had 4 judicial IRV elections in 2010 - 3 for county superior court, and one statewide. Of the 4, two (including the one statewide race) went to IRV for final tabulation. In both of those races, the 1st round winner was not the winner in the final round - they both flipped! In the statewide race, Judge Thigpen went from being 100K votes up to being 6.6K votes down - that's 2.4 votes per precinct across the state! But it's odd that NC's earlier IRV election was won by the 1st place winner - like about 98% of IRV races around the country! In 2010, those odds were totally flipped - 100% of our IRV races flipped, but we can't be sure if they were counted correctly. Too many variables.

That's a good reason why IRV fails as a voting method - it's too tough to count. When you add it to the evidence that it's too complicated and confusing for many voters to use properly, and it's too expensive when you honestly and accurately account for all the costs. There are no certified IRV voting systems because there are so many different types of IRV elections and it's too expensive for vendors to run them through the federal labs for testing.

No - I'd rather see traditional primary elections with runoffs if needed. And why have different primaries at different times for presidential elections? We have 100 counties in NC - it makes no sense to start the race for governor in two small counties and roll them across the state over a couple of months just for tradition's sake.

With the internet and other modern news media, all campaigning is nationwide. Let's cut the bull - start the campaigns in January, have a one-day national primary in May, then do the conventions in July and start the national election campaigns in August. IRV is a waste of time and money and it will only confuse voters in DC! And if you want to do a test - don't try and use trickery to make it seem easy. Elections for candidates have real-life consequences, so don't make a game out of it by trying to trick people with ice cream or other desert flavors!

by Chris Telesca on Feb 2, 2011 12:53 pm • link • report

If IRV is too difficult for our election administrators, that's a tragic statement about the state of our democracy.

by David desJardins on Feb 2, 2011 12:57 pm • link • report

@Chris Telesca: Don't voters deserve to know which candidate follows the platform of their preferred party? And don't those same voters deserve to know which of their party's candidates deserve their support after winning a primary election?

I do think that part of the argument for IRV is to encourage voters to support candidates rather than parties. People who think that voting party preference is better than voting candidate preference, and who like the fact that partisans who turn out in primaries have more effect on the outcome in a primary+general election system than the less partisan voters who vote only in general elections, are going to prefer the present system.

by David desJardins on Feb 2, 2011 1:02 pm • link • report

Wow, David, you're sure right about the IRV people coming out of the wordwork! I do want to respond to a more general comment from @Chris Telesca about open primaries though:

Don't voters deserve to know which candidate follows the platform of their preferred party? And don't those same voters deserve to know which of their party's candidates deserve their support after winning a primary election?

I'm not sure about the party structure in places like North Carolina, but this sort of logic tends not to work in places like DC where one of the two political parties is extremely dominant. What happens in practice is that a relatively low-turnout primary anoints a nominee (for the Democrats in DC, though I'm sure it's true with Republicans elsewhere) that essentially cannot lose in the general election. Thus, the real election is the battle for the votes of those who are partisan enough and politically interested enough to vote in the primary; this does not necessarily represent the will of the electorate at large, even in one-party-dominant jurisdictions like DC.

I have no problem with open elections where everyone's on the same ballot but their party affiliation is listed next to their name. I just think that the general electorate as a whole ought to get to choose among all the candidates. There are often substantive differences even among those within the same political party.

by jfruh on Feb 2, 2011 1:07 pm • link • report

I like IRV voting because voters usually can easily identify their preference(s)[when more than two candidates are on the ballot]. If ONE DC election were held instead of separate primaries and a general election each cycle, how would partian contests, ie: election of Democrats and election of Republicans to their party central/state committees be deterimined usung IRV. And how would IRV be used to select national convention delegates to the DNC and RNC?

by curious mem on Feb 2, 2011 1:08 pm • link • report

IRV is the right way to go, but the entrenched interests (sad to say, the Dems) will never go for anything that may dilute their power.

by drez on Feb 2, 2011 2:22 pm • link • report

@ David desJardins If you vote for both Alice and Bob, then your vote would be wasted and irrelevant to the outcome.

I'm not sure why voting for two candidates you approve of, and having one of them win counts as "wasting" your vote. If I felt both Alice and Bob were capable candidates I'd vote for them.

Do you think IRV can't result in strategic voting? No voting system is immune to strategic voting. IRV is better than the current plurality system, but it still is far from optimal. In the last Burlington mayoral election that used IRV (they repealed it next cycle), there wound up being a spoiler and the race was won by a candidate who was not the preference of the most voters.

by jcm on Feb 2, 2011 6:29 pm • link • report

@jcm: I'm not sure why voting for two candidates you approve of, and having one of them win counts as "wasting" your vote.

It's because people want their input to affect the election. Are you seriously suggesting that if I think every candidate will do an acceptable job, that I shouldn't bother to vote at all, since I "approve of" every single candidate then my vote shouldn't count toward getting a candidate that I prefer over one that I don't??

In the last Burlington mayoral election that used IRV (they repealed it next cycle), there wound up being a spoiler and the race was won by a candidate who was not the preference of the most voters.

I don't know what you mean by "the preference of the most voters". It's entirely possible for a majority voters to prefer A to B, a majority to prefer B to C, and a majority to prefer C to A. Yet someone has to win.

by David desJardins on Feb 2, 2011 8:23 pm • link • report

Are you seriously suggesting that if I think every candidate will do an acceptable job, that I shouldn't bother to vote at all, since I "approve of" every single candidate then my vote shouldn't count toward getting a candidate that I prefer over one that I don't??
No, I'm not suggesting that at all.

I don't know what you mean by "the preference of the most voters". It's entirely possible for a majority voters to prefer A to B, a majority to prefer B to C, and a majority to prefer C to A. Yet someone has to win.

No, that's not what happened. Here's a very detailed look. The short version:

Montroll was preferred vs Kiss by a voter majority, margin 590.
Montroll was preferred vs Wright by a voter majority, margin 929.
Montroll was preferred vs Smith by a voter majority, margin 1575.
Montroll was preferred vs Simpson by a voter majority, margin 5676.
Montroll was preferred vs All Write-Ins Combined by a voter majority, margin 6554.
There were no other candidates. IRV elected Kiss.

by jcm on Feb 2, 2011 9:02 pm • link • report

So, the short version is that IRV worked better than plurality voting in Burlington, and no one knows what would have happened under approval or range voting. That doesn't sound like much of a case against IRV, while I still think the case against approval voting is clearcut (and range voting is truly awful, each voter needs to do a huge amount of analysis just to think about how to vote).

by David desJardins on Feb 2, 2011 9:21 pm • link • report

We use approval voting for our homeowners association. Basically because we don't care how many people are on the board. It's usually more of a problem getting people to run (we need about a five member board for 56 townhouses) than it is to select just one from many candidates.

Any person getting more than 50% approval is elected. Our old rules were terrible. They approved a change that had the side effect of allowing anyone with even one vote to be elected.

Ours is kind of a special case.

by Michael Perkins on Feb 2, 2011 9:38 pm • link • report

No, the short version is that IRV failed. Did you miss this part?

Kiss won, but if 753 Wright-voters had switched their vote to Kiss, that would have made Kiss lose.

I don't understand your clear-cut case against approval voting at all. I suspect we're unlikely to convince each other, though, so I think we'll just have to agree to disagree.

by jcm on Feb 2, 2011 9:41 pm • link • report

Did you miss this part?

No. Did you miss the part where plurality voting would have done worse, and no one knows how approval voting would have done?

The straightforward case against approval voting is that in IRV it's usually effective to just vote your preferences, while in approval voting it's very hard to figure out how to vote. Deciding whom to "approve" when I prefer A to B, but B to C, is very complicated. Most voters don't think they have some candidates who are all equally good, and other candidates who are all equally bad. They have preferences between them.

by David desJardins on Feb 2, 2011 9:45 pm • link • report

In the off chance that you want to improve policy rather than merely theorize, I suggest you look at the approach Louisiana and North Carolina took. Rather than move their primaries into the summer, they added the types of innovative voting discussed here for the overseas ballots only. You could make a good case that the administrative nightmare of a new voting system for overseas ballots only would be less than the nightmare of moving the voting into the summer.

by JimT on Feb 3, 2011 8:52 am • link • report