"It's too confusing" is what you say if you oppose it for other reasons that you don't want to identify. Like: it weakens the power of party extremists. Open primaries combined with instant run-off elections shift power from party extremists to the overall electorate. By creating a pull towards the center, it helps create coalitions that can effectively govern. The importance of this reform should not be dismissed as too-small-to-be-worth-it; over time, the effect on our political climate of increasing the reward of centrism and reducing the rewards of extremism will be huge. These are reforms that America needs at this point in time. (I agree with Matt's ambiguous feelings about semi-open primaries; much better to do it right. And it's a fair question whether at-large city councilors are better than regional ones, but it doesn't bear on the argument for open primaries + RCV, especially nationally.)
While it is slightly more complicated in terms of voter education and ballot construction, I like cumulative voting as an option for multi-candidate races. It does require that factions correctly match their slates to their electoral strength, but that is within their control.
Cumulative Voting and Single Transferrable Voting are particularly good for electing planning groups, where recommendations from a group that has broad representation is much more likely to be respected by a City Council than a group that only includes the views of one segment of the community. (Winning by sharing.)
It is also better than having designated seats for given segments because these voting systems automatically adapt to the introduction of new factions or changes to old factions.
Regarding RCV, I originally didn't think that it mattered in California because we have an open primaries and one-party rule, so in theory the two most popular Democrats make it to the runoff.
In practice, though, the party elites decide to put their thumb on the scale for one candidate and then run a false campaign to promote the perceived weakest candidate (usually a Republican) in the race to prevent the most viable alternative candidate from making it to the runoff. This is why moderate Democrats in California wish for a viable second party, even if it means having the Republican Party become relevant again.
Strong disagree with Democrats who think the key to a better election system is forcing the least informed and motivated people to vote. People who don't care shouldn't vote.
I've heard lots of people say it, none of them were Republicans, though I don't know if I have ever heard an elected Democrat talk about it. In general, if you hear about a get-out-the-vote program, chances are the people who started it are Democrats. If people are fighting for expanding voter access, chances are it's a Democrat doing it. If someone is trying to restrict voting access under the name of fraud prevention, it's always a Republican. I think it reflects well on Democrats that they want to maximize ballot access, but I think they are wrong to think it will improve election outcomes.
I just voted in the first election since my city (Portland, Oregon) adopted RCV, and am very curious what the electoral results will be. However, one thing that affected the voting experience this cycle that may be related is there were 20+ candidates to choose from (no primary?) the result was choice overload. The average voter will not have the time to carefully weigh the top 6 out of 20 candidates after researching all of them. This will lead to arbitrary voting, though in fairness, maybe that happens already in low-name-ID local elections.
"This will lead to arbitrary voting, though in fairness, maybe that happens already in low-name-ID local elections."
I'm very sure that already happens in low-name-ID local elections because I'm probably in the 90th percentile or higher of Americans for political engagement + education and I end up voting arbitrarily for some local races because there are always a few candidates for several offices who somehow can't be bothered to put any sort of information about themselves on-line or even respond to newspaper "candidate forums" or anything like that.
Has Matt said what he doesn’t like about councilmember Robert White? I have come around to thinking he would have been better as mayor than Bowser getting a 3rd term. Who is a better option on the ballot than Robert White?
He was talking about a proportional representation system with 13 seats. In a truly proportional representation system, you would only need 7.7% of the vote to get one seat.
Pre-Donald Trump, there were Republican presidential candidates who got ~7% in DC. Someone running for local office should be able to do better than that since they could tailor their campaign message to local concerns and ignore stuff like cutting federal spending, abortion, and gun control.
I live in Oakland, Ca, which has used RCV for more than a decade. The last mayoral election was pretty contentious and close. There definitely was a lot of noise questioning the legitimacy of the the RCV results. I think some of this was (unintentional?) misinformation. I saw some people on Nextdoor advocating only writing down one candidate to make sure "your vote doesn't go to other candidates". (RCV nerds were quick to respond in the comments, but still.) Others were complaining that the relatively complicated reasoning whose Nth choice vote goes to which candidate was actually subterfuge hiding a stolen election. It's hard to know how good faith these arguments were... I suspect the 2016 and 2020 presidential elections have reduced trust in the electoral system at all levels (on the left and right respectively). It's also quite plausible that a FPTP election would have been perceived as even less legitimate.
Still, there might be some opportunity for standardized visualizations or interactive web tools showing how different ballots counted ("your first choice was eliminated in the first round...").
"It's also quite plausible that a FPTP election would have been perceived as even less legitimate."
Even knowing nothing about local Oakland politics, I'm deeply skeptical about that. That's because I think FPTP inherently possesses greater legitimacy with the general public due to its transparency and simplicity. I mean, one thing I really can't help noticing, is that if you look *outside* the realm of official elections -- i.e., where everyday people are actually free to make their own decisions about voting structures without needing to consider anything about what is constitutional or what incumbents might approve, etc. -- FPTP is for all practical purpose the exclusive form of voting used, with the only variation being whether a simple plurality is enough or if there needs to be a runoff if there's no majority first pick. To put it another way, when the family is choosing where to go for dinner or friends are picking what movie to go to or the senior class has to decide on a prom theme or the firm is deciding which of three venues to hold the office picnic at, etc., FPTP is what invariably gets used.
The most comprehensive analysis to-date found 5% of the votes for ranked races in the US showing some kind of problem, 10x higher than on non-ranked races _on the same ballot_. (To be fair, _some_ of these mismarks probably get fixed by the system -- e.g. if you skip the first rank, but then do a consistent series of 2nd, 3rd, 4th. But lots of them get tossed, easily 3+%.)
Approval ballots would avoid this problem. This would be fairer, and would cost much less (both in actual printing and tabulation costs, and voter education efforts). Basic Approval voting can be implemented, _for free_, with any equipment that's capable of running "vote for N" type races, or Yes/No ballot propositions. (You can treat an Approval race as a vote for N, where N is always the same as the number of candidates running. Or you can treat each candidate like a Y/N choice.)
In Portland, OR, RCV has resulted in a city council race with 98 candidates. In some districts voters will have 30 candidates to evaluate. How many voters are going to be able to conduct enough research to understand who all these candidates are? Now imagine a big, sloppy, chaotic candidate pool for every race.
If this is what RCV looks like, I am not hopeful that it will lead to better outcomes.
One thing I would like to see done with RCV is improvements to the counting process to avoid long delays in results. The way RCV elections are typically administered, they release a running total of the first round and keep all the 2nd/3rd choices secret until every single vote has been counted (including a few mail-in votes that arrived days late). And even then, they only release what's strictly necessary to determine the winner. For example, in the Maine 2020 Senate election, we have no idea how close Gideon actually came to beating Collins because, with Collins just barely clearing the 50% mark in the first round, Maine never released how many votes for the 3rd candidate (a progressive) chose Gideon as the 2nd choice vs. somehow else or blank. Yes, it doesn't matter for purposes of determining the winner, but the public should still have a right to know how close the race actually was, just like, with First Past the Post, the state doesn't just stop counting votes, once a winning candidate has 50% + 1 of all votes received.
Ideally, I think the way it would work, the state would publish two views of the running results, a summary view and a detailed view. The summary view might show only the first round total and, maybe, what the rest of the rounds would look like if the current total were final. The detailed view would show vote counts for all possible permutations of candidates (with 4 candidates, there are only 24 of them) and could answer questions such as how the race between two major candidates would be affected, depending on which minor candidate gets eliminated first. While many people might get overwhelmed by all the numbers in the detailed view, that's what the summary view would be for, the detailed view would be intended to be used primarily by media outlets who could put together their own digestible summary that might show a little bit more information than what the summary view shows directly.
I suspect what may actually be going on is that the vote-counting software they're using was designed for first-past-the-post, so gathering anything beyond the first choice requires running all the ballots through the machines additional times, plus a bunch of manual arithmetic. But, that's just a relic of how the tabulation machines and the software are designed, it's not inherent in how ranked-choice voting has to be.
"just like, with First Past the Post, the state doesn't just stop counting votes, once a winning candidate has 50% + 1 of all votes received."
My understanding is that some jurisdictions at least used to actually do just that. I've seen stuff about it from the Libertarian Party because they fight to have every vote counted in races that they are in as for some jurisdictions hitting particular numerical thresholds can be important for third-party ballot access in subsequent elections.
"It's too confusing" is what you say if you oppose it for other reasons that you don't want to identify. Like: it weakens the power of party extremists. Open primaries combined with instant run-off elections shift power from party extremists to the overall electorate. By creating a pull towards the center, it helps create coalitions that can effectively govern. The importance of this reform should not be dismissed as too-small-to-be-worth-it; over time, the effect on our political climate of increasing the reward of centrism and reducing the rewards of extremism will be huge. These are reforms that America needs at this point in time. (I agree with Matt's ambiguous feelings about semi-open primaries; much better to do it right. And it's a fair question whether at-large city councilors are better than regional ones, but it doesn't bear on the argument for open primaries + RCV, especially nationally.)
While it is slightly more complicated in terms of voter education and ballot construction, I like cumulative voting as an option for multi-candidate races. It does require that factions correctly match their slates to their electoral strength, but that is within their control.
Cumulative Voting and Single Transferrable Voting are particularly good for electing planning groups, where recommendations from a group that has broad representation is much more likely to be respected by a City Council than a group that only includes the views of one segment of the community. (Winning by sharing.)
It is also better than having designated seats for given segments because these voting systems automatically adapt to the introduction of new factions or changes to old factions.
Regarding RCV, I originally didn't think that it mattered in California because we have an open primaries and one-party rule, so in theory the two most popular Democrats make it to the runoff.
In practice, though, the party elites decide to put their thumb on the scale for one candidate and then run a false campaign to promote the perceived weakest candidate (usually a Republican) in the race to prevent the most viable alternative candidate from making it to the runoff. This is why moderate Democrats in California wish for a viable second party, even if it means having the Republican Party become relevant again.
Sorry: Is London Breed a person, or has San Francisco elected a show dog as its mayor?
no one is even going to comment that Matt apparently thinks there are only 43 states in the union?
Strong disagree with Democrats who think the key to a better election system is forcing the least informed and motivated people to vote. People who don't care shouldn't vote.
Which Democrats say we should have mandatory voting (like Brazil or Australia)
I've heard lots of people say it, none of them were Republicans, though I don't know if I have ever heard an elected Democrat talk about it. In general, if you hear about a get-out-the-vote program, chances are the people who started it are Democrats. If people are fighting for expanding voter access, chances are it's a Democrat doing it. If someone is trying to restrict voting access under the name of fraud prevention, it's always a Republican. I think it reflects well on Democrats that they want to maximize ballot access, but I think they are wrong to think it will improve election outcomes.
I just voted in the first election since my city (Portland, Oregon) adopted RCV, and am very curious what the electoral results will be. However, one thing that affected the voting experience this cycle that may be related is there were 20+ candidates to choose from (no primary?) the result was choice overload. The average voter will not have the time to carefully weigh the top 6 out of 20 candidates after researching all of them. This will lead to arbitrary voting, though in fairness, maybe that happens already in low-name-ID local elections.
"This will lead to arbitrary voting, though in fairness, maybe that happens already in low-name-ID local elections."
I'm very sure that already happens in low-name-ID local elections because I'm probably in the 90th percentile or higher of Americans for political engagement + education and I end up voting arbitrarily for some local races because there are always a few candidates for several offices who somehow can't be bothered to put any sort of information about themselves on-line or even respond to newspaper "candidate forums" or anything like that.
Matt, I'm surprised you didn't mention the Voting Rights Act. PR would be illegal in DC due to the VRA and its effect on Black voter power, right?
Has Matt said what he doesn’t like about councilmember Robert White? I have come around to thinking he would have been better as mayor than Bowser getting a 3rd term. Who is a better option on the ballot than Robert White?
My ideal DC Council reform is having an equal number of ward and at-large CMs with wards using RCV and at-large using PR
You think a Republican could be elected to DC City Council? hahahaha
He was talking about a proportional representation system with 13 seats. In a truly proportional representation system, you would only need 7.7% of the vote to get one seat.
Ok, but do you really think any Republican gets to 7% of the vote in DC? I wouldn’t expect they could top even 5%…
Pre-Donald Trump, there were Republican presidential candidates who got ~7% in DC. Someone running for local office should be able to do better than that since they could tailor their campaign message to local concerns and ignore stuff like cutting federal spending, abortion, and gun control.
I live in Oakland, Ca, which has used RCV for more than a decade. The last mayoral election was pretty contentious and close. There definitely was a lot of noise questioning the legitimacy of the the RCV results. I think some of this was (unintentional?) misinformation. I saw some people on Nextdoor advocating only writing down one candidate to make sure "your vote doesn't go to other candidates". (RCV nerds were quick to respond in the comments, but still.) Others were complaining that the relatively complicated reasoning whose Nth choice vote goes to which candidate was actually subterfuge hiding a stolen election. It's hard to know how good faith these arguments were... I suspect the 2016 and 2020 presidential elections have reduced trust in the electoral system at all levels (on the left and right respectively). It's also quite plausible that a FPTP election would have been perceived as even less legitimate.
Still, there might be some opportunity for standardized visualizations or interactive web tools showing how different ballots counted ("your first choice was eliminated in the first round...").
"It's also quite plausible that a FPTP election would have been perceived as even less legitimate."
Even knowing nothing about local Oakland politics, I'm deeply skeptical about that. That's because I think FPTP inherently possesses greater legitimacy with the general public due to its transparency and simplicity. I mean, one thing I really can't help noticing, is that if you look *outside* the realm of official elections -- i.e., where everyday people are actually free to make their own decisions about voting structures without needing to consider anything about what is constitutional or what incumbents might approve, etc. -- FPTP is for all practical purpose the exclusive form of voting used, with the only variation being whether a simple plurality is enough or if there needs to be a runoff if there's no majority first pick. To put it another way, when the family is choosing where to go for dinner or friends are picking what movie to go to or the senior class has to decide on a prom theme or the firm is deciding which of three venues to hold the office picnic at, etc., FPTP is what invariably gets used.
Just as a factual matter, the rate of mis-marking on ranked ballots is _shockingly_ high.
https://www.youtube.com/watch?v=DdPIauxbTtI
The most comprehensive analysis to-date found 5% of the votes for ranked races in the US showing some kind of problem, 10x higher than on non-ranked races _on the same ballot_. (To be fair, _some_ of these mismarks probably get fixed by the system -- e.g. if you skip the first rank, but then do a consistent series of 2nd, 3rd, 4th. But lots of them get tossed, easily 3+%.)
Approval ballots would avoid this problem. This would be fairer, and would cost much less (both in actual printing and tabulation costs, and voter education efforts). Basic Approval voting can be implemented, _for free_, with any equipment that's capable of running "vote for N" type races, or Yes/No ballot propositions. (You can treat an Approval race as a vote for N, where N is always the same as the number of candidates running. Or you can treat each candidate like a Y/N choice.)
IxNay on proportional representation. Italy has it and it can't vote it's way out of brown paper bag. Israel has it, 'nuff said. No thank you.
I'm all for the states being our laboratory of democracy. Let's get going with the experiments.
In Portland, OR, RCV has resulted in a city council race with 98 candidates. In some districts voters will have 30 candidates to evaluate. How many voters are going to be able to conduct enough research to understand who all these candidates are? Now imagine a big, sloppy, chaotic candidate pool for every race.
If this is what RCV looks like, I am not hopeful that it will lead to better outcomes.
One thing I would like to see done with RCV is improvements to the counting process to avoid long delays in results. The way RCV elections are typically administered, they release a running total of the first round and keep all the 2nd/3rd choices secret until every single vote has been counted (including a few mail-in votes that arrived days late). And even then, they only release what's strictly necessary to determine the winner. For example, in the Maine 2020 Senate election, we have no idea how close Gideon actually came to beating Collins because, with Collins just barely clearing the 50% mark in the first round, Maine never released how many votes for the 3rd candidate (a progressive) chose Gideon as the 2nd choice vs. somehow else or blank. Yes, it doesn't matter for purposes of determining the winner, but the public should still have a right to know how close the race actually was, just like, with First Past the Post, the state doesn't just stop counting votes, once a winning candidate has 50% + 1 of all votes received.
Ideally, I think the way it would work, the state would publish two views of the running results, a summary view and a detailed view. The summary view might show only the first round total and, maybe, what the rest of the rounds would look like if the current total were final. The detailed view would show vote counts for all possible permutations of candidates (with 4 candidates, there are only 24 of them) and could answer questions such as how the race between two major candidates would be affected, depending on which minor candidate gets eliminated first. While many people might get overwhelmed by all the numbers in the detailed view, that's what the summary view would be for, the detailed view would be intended to be used primarily by media outlets who could put together their own digestible summary that might show a little bit more information than what the summary view shows directly.
I suspect what may actually be going on is that the vote-counting software they're using was designed for first-past-the-post, so gathering anything beyond the first choice requires running all the ballots through the machines additional times, plus a bunch of manual arithmetic. But, that's just a relic of how the tabulation machines and the software are designed, it's not inherent in how ranked-choice voting has to be.
"just like, with First Past the Post, the state doesn't just stop counting votes, once a winning candidate has 50% + 1 of all votes received."
My understanding is that some jurisdictions at least used to actually do just that. I've seen stuff about it from the Libertarian Party because they fight to have every vote counted in races that they are in as for some jurisdictions hitting particular numerical thresholds can be important for third-party ballot access in subsequent elections.