Steven, we disagree on the *fundamental* undesirability of plurality seat allocation. This shapes other features of our perspectives. One might be how we view a "wasted vote." Another might be the value of "winner-take-all" as a descriptor. I think this disagreement is producing the debate you want to have on this blog.
As for "vote leakage," I hypothesized and concluded from evidence that this was a proximate cause of STV repeal in three key U.S. cases. I say "key" because those cases shaped the thinking of the national PR lobby. In other cases, we find people making arguments about STV that suggest frustration with "leakage." Leakage may present other problems, but it was central to my answer to the research question: "Why did STV get repealed after having been adopted in U.S. cities?"
This is an interesting defense of (P)RCV in comparison to Party List PR. Your piece offers a lot to think about regarding applications of Party List PR.
What gives me pause is that even in the Australian Senate which basically uses PRCV, there’s still a dominant two-party system even though it is less stark than in the Australian House which uses IRV. In both chambers, I’m not seeing the sheer diversity of parties that I see in countries with Party List or MMPR countries. If applied feature for feature for legislatures in a presidential republic like ours, how would third parties like the Greens and Libertarians perform in terms of winning seats?
Plus, in Australia, they’ve mostly done away with group ticket voting, the RCV equivalent of the straight ticket device (IMO), and have switched to allowing voters to cast less than the full number of rankings on the ballot, allowing for voters to exhaust their ballots early if their candidate doesn’t make it to the next round of counting. I don’t know if you addressed this in the post, but doesn’t that also count as wasted votes? Not to mention, again, the minimal third-party candidate success in single-winner IRV.
Also, If it’s ok, I’d also like to see the wasted vote figures in presidential republics which use Party List PR for legislative elections like Brazil and most of Central and South America. They also have a history of a broad range of parties being elected to the same chambers, as well as negotiations for building pro-government and opposition coalitions, except to support the president rather than a prime minister. Since there are no presidential republics which use RCV at the national level to elect all members of a lower or unicameral chamber, I think it would help to compare elections in other presidential republics’ legislatures to ours as well.
But your post is one of the few times I’ve seen anyone go to bat in favor of RCV vis-a-vis Party List PR, so I do appreciate the perspective.
While I respect that there are some issues with OLPR, I can't help but feel this is a bit of a strawman. Many of the issues being alleged here apply equally well to other proportional systems. Let me go through some of these claims.
> This feature unleashes a second level of competition – a high degree of intra-party competition, because within each party the candidates that win seats are those that garner the most personal votes from voters.
True, but in STV elections there is also intraparty competition where only the most popular candidates from each party will win seats. The only way to avoid candidates competing against those in their party while maintaining proportionality is to use a closed-list system, but that just moves the competition to behind closed doors. In other words, this is an unavoidable problem.
The competition takes on different flavors in these systems. In STV, voters choose where their vote transfers while in OLPR votes pool with everyone from the same party. Which of these makes the intraparty competition worse? I honestly don't know.
> Because candidates within the same political party are running against each other, that means they can spoil each other’s chance of election. Spoiler candidates within a political party can undermine the intra-party solidarity between those candidates.
This is true. But there are many ways to set up a List system, some of which deal with this problem better than others. In fact, I would argue the Finnish system isn't ideal. Most of the [open list systems](https://en.wikipedia.org/wiki/Open_list) being used have some kind of threshold vote that must be met for a candidate to "jump the line" -- otherwise candidates are elected in party list order. A moderate threshold like 5% assures that candidates voters want are elected, but that candidates with a "marginal share of the vote" don't get a free ride to office like you described. Another popular solution is [panachage](https://en.wikipedia.org/wiki/Panachage), where voters can vote for as many candidates as there are seats, which also avoids the popular candidate problem. In fact, voters can even split their votes between different parties. Arguably, OLPR and STV don't have to be exclusive, and you could use STV to determine party winners while allocating party seats proportionally.
> Or imagine a scenario in which, within a political party, there are two factions, one more conservative than the other.
Hmm, why not form separate parties in this case? Heck, with an alliance system they could still pool their votes together like they were one party so this wouldn't cost them any seats. In fact, this is arguably the point. OLPR proponent Shugart [claims](https://fruitsandvotes.wordpress.com/2021/01/19/emergency-electoral-reform-olpr-for-the-us-house/) "the overriding objective is to let different wings of current parties compete separately."
> In the one district of Uusimaa, the total number of fielded candidates in one election amounted to 395, and each voter was required to pick a single candidate.
[Uusimaa](https://en.wikipedia.org/wiki/Uusimaa_(parliamentary_electoral_district)) elects 37 candidates to parliament, the largest in Finland. If there were just two parties we would already expect 74 names on the ballot. But larger districts also increases the number of political parties that can potentially win, so the number of running candidates increases superlinearly, leading to a crazy high number like 395. We would expect this result in both STV and OLPR. In Australia, the New South Wales election for Senate seats had 75 candidates competing for just six seats. The difference between the systems is that in OLPR you have to pick one candidate that is your favorite, while in STV you need to rank all 75 candidates (or fewer and risk wasting your vote). Since ranking necessarily involves picking your favorite, STV requires strictly more work from the voters when it comes to managing the election of large magnitude districts.
> What’s the best way for a candidate to cultivate their personal vote? The same way you do in the US-style plurality system – name recognition.
The same is true for candidates to try to win votes in STV. Again, this isn't the fault of OLPR here.
> The “wasted vote” problem in proportional methods
This I mostly agree with. For some context though, the 2016 Australian election had 7.52% of votes exhaust and 49.39% of ballots ([source](https://www.aec.gov.au/About_AEC/research/files/sbps-exhaustion.pdf). The reason the number of ballots exhausted is so much higher is because many of the ballots are being exhausted after using some of their vote to elect a candidate. So far more ballots end up wasting a small amount of their vote under STV, but in OLPR a smaller amount of ballots waste all of their vote.
> Benjamin Netanyahu, achieved only 23 percent of the popular vote.
Overall I probably lean on your side in terms of STV v OLPR. But I think you undersell OLPR here, and that the system works a bit better than how you describe it.
Steven, we disagree on the *fundamental* undesirability of plurality seat allocation. This shapes other features of our perspectives. One might be how we view a "wasted vote." Another might be the value of "winner-take-all" as a descriptor. I think this disagreement is producing the debate you want to have on this blog.
As for "vote leakage," I hypothesized and concluded from evidence that this was a proximate cause of STV repeal in three key U.S. cases. I say "key" because those cases shaped the thinking of the national PR lobby. In other cases, we find people making arguments about STV that suggest frustration with "leakage." Leakage may present other problems, but it was central to my answer to the research question: "Why did STV get repealed after having been adopted in U.S. cities?"
Thank you for sharing the paper by Von Schoultz.
This is an interesting defense of (P)RCV in comparison to Party List PR. Your piece offers a lot to think about regarding applications of Party List PR.
What gives me pause is that even in the Australian Senate which basically uses PRCV, there’s still a dominant two-party system even though it is less stark than in the Australian House which uses IRV. In both chambers, I’m not seeing the sheer diversity of parties that I see in countries with Party List or MMPR countries. If applied feature for feature for legislatures in a presidential republic like ours, how would third parties like the Greens and Libertarians perform in terms of winning seats?
Plus, in Australia, they’ve mostly done away with group ticket voting, the RCV equivalent of the straight ticket device (IMO), and have switched to allowing voters to cast less than the full number of rankings on the ballot, allowing for voters to exhaust their ballots early if their candidate doesn’t make it to the next round of counting. I don’t know if you addressed this in the post, but doesn’t that also count as wasted votes? Not to mention, again, the minimal third-party candidate success in single-winner IRV.
Also, If it’s ok, I’d also like to see the wasted vote figures in presidential republics which use Party List PR for legislative elections like Brazil and most of Central and South America. They also have a history of a broad range of parties being elected to the same chambers, as well as negotiations for building pro-government and opposition coalitions, except to support the president rather than a prime minister. Since there are no presidential republics which use RCV at the national level to elect all members of a lower or unicameral chamber, I think it would help to compare elections in other presidential republics’ legislatures to ours as well.
But your post is one of the few times I’ve seen anyone go to bat in favor of RCV vis-a-vis Party List PR, so I do appreciate the perspective.
While I respect that there are some issues with OLPR, I can't help but feel this is a bit of a strawman. Many of the issues being alleged here apply equally well to other proportional systems. Let me go through some of these claims.
> This feature unleashes a second level of competition – a high degree of intra-party competition, because within each party the candidates that win seats are those that garner the most personal votes from voters.
True, but in STV elections there is also intraparty competition where only the most popular candidates from each party will win seats. The only way to avoid candidates competing against those in their party while maintaining proportionality is to use a closed-list system, but that just moves the competition to behind closed doors. In other words, this is an unavoidable problem.
The competition takes on different flavors in these systems. In STV, voters choose where their vote transfers while in OLPR votes pool with everyone from the same party. Which of these makes the intraparty competition worse? I honestly don't know.
> Because candidates within the same political party are running against each other, that means they can spoil each other’s chance of election. Spoiler candidates within a political party can undermine the intra-party solidarity between those candidates.
This is true. But there are many ways to set up a List system, some of which deal with this problem better than others. In fact, I would argue the Finnish system isn't ideal. Most of the [open list systems](https://en.wikipedia.org/wiki/Open_list) being used have some kind of threshold vote that must be met for a candidate to "jump the line" -- otherwise candidates are elected in party list order. A moderate threshold like 5% assures that candidates voters want are elected, but that candidates with a "marginal share of the vote" don't get a free ride to office like you described. Another popular solution is [panachage](https://en.wikipedia.org/wiki/Panachage), where voters can vote for as many candidates as there are seats, which also avoids the popular candidate problem. In fact, voters can even split their votes between different parties. Arguably, OLPR and STV don't have to be exclusive, and you could use STV to determine party winners while allocating party seats proportionally.
> Or imagine a scenario in which, within a political party, there are two factions, one more conservative than the other.
Hmm, why not form separate parties in this case? Heck, with an alliance system they could still pool their votes together like they were one party so this wouldn't cost them any seats. In fact, this is arguably the point. OLPR proponent Shugart [claims](https://fruitsandvotes.wordpress.com/2021/01/19/emergency-electoral-reform-olpr-for-the-us-house/) "the overriding objective is to let different wings of current parties compete separately."
> In the one district of Uusimaa, the total number of fielded candidates in one election amounted to 395, and each voter was required to pick a single candidate.
[Uusimaa](https://en.wikipedia.org/wiki/Uusimaa_(parliamentary_electoral_district)) elects 37 candidates to parliament, the largest in Finland. If there were just two parties we would already expect 74 names on the ballot. But larger districts also increases the number of political parties that can potentially win, so the number of running candidates increases superlinearly, leading to a crazy high number like 395. We would expect this result in both STV and OLPR. In Australia, the New South Wales election for Senate seats had 75 candidates competing for just six seats. The difference between the systems is that in OLPR you have to pick one candidate that is your favorite, while in STV you need to rank all 75 candidates (or fewer and risk wasting your vote). Since ranking necessarily involves picking your favorite, STV requires strictly more work from the voters when it comes to managing the election of large magnitude districts.
> What’s the best way for a candidate to cultivate their personal vote? The same way you do in the US-style plurality system – name recognition.
The same is true for candidates to try to win votes in STV. Again, this isn't the fault of OLPR here.
> The “wasted vote” problem in proportional methods
This I mostly agree with. For some context though, the 2016 Australian election had 7.52% of votes exhaust and 49.39% of ballots ([source](https://www.aec.gov.au/About_AEC/research/files/sbps-exhaustion.pdf). The reason the number of ballots exhausted is so much higher is because many of the ballots are being exhausted after using some of their vote to elect a candidate. So far more ballots end up wasting a small amount of their vote under STV, but in OLPR a smaller amount of ballots waste all of their vote.
> Benjamin Netanyahu, achieved only 23 percent of the popular vote.
This is misleading. While his party may have only got 23%, his coalition [got 48.1%](https://fruitsandvotes.wordpress.com/2022/11/02/coordination-failure-under-nationwide-pr-manufactured-majority-in-israel-2022/). Jumping from 48% to a majority isn't great, but not nearly as bad as going from 23% to a majority like you imply.
------
Overall I probably lean on your side in terms of STV v OLPR. But I think you undersell OLPR here, and that the system works a bit better than how you describe it.