IRV's increase in the candidate pool size dissipates after several election cycles
I have seen this claimed in the past by IRV detractors. Looks like there is possibly bite to that bark --- title is quoted from the abstract of this paper https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4317298
This research project investigates whether one of the purported benefits of Ranked Choice Voting (RCV) – an expanded candidate pool – has been realized in the cities where it has been implemented. Employing a difference-in-differences approach with matching, I find that any increase in the candidate pool size dissipates after several election cycles. Similarly, related benefits such as a higher quality and more diverse candidate pool are also temporary. As the first project to systematically examine whether and how RCV influences candidate entry in the US, this analysis questions some of the purported benefits of RCV as well as offers a comprehensive new resource for scholars and practitioners investigating the effects of RCV.
Another reason we need proportional representation! the only way to get true candidate variety.
Sass last edited by
@andy-dienes What we need are voting methods that are stable with more than two competitive candidates, and RCV is not one of them. We haven't seen dissipation in Fargo or St. Louis, which have both had two Approval elections. Granted, STL halved their number of wards in this last round, so I wouldn't call the evidence definitive, but PR is not the only way to get true candidate variety.
but PR is not the only way to get true candidate variety.
Maybe. very much remains to be seen --- a grand total of three primaries across two cities is not even close to enough to draw any real conclusions for approval. that paper I linked above collected "data on over six thousand unique elections occurring between 1995 and 2022 across 273 different cities" and analyzes 45 treatment (i.e. IRV) cities, and even then the author admits that the sample size is very small.
meanwhile there is no shortage of strong evidence that PR rules increase candidate pool size and diversity.
@andy-dienes said in IRV's increase in the candidate pool size dissipates after several election cycles:
the only way to get true candidate variety
This should be framed as a positive, not a negative. I very much doubt that it's the "only way" but I'd strongly agree if you had said it's "a proven way".
The question is: why is there still a small candidate pool with IRV? Is it just a thing you will always get with single-winner elections? I don't think you get massively different election results under Condorcet (which is why the single example of Burlington 2009 keeps coming up), so there's probably not much reason to think it would change the number of candidates by much either. Then you have cardinal methods (e.g. approval, score, STAR) - but is there any particular reason why these would make much difference compared to ranked methods?
Is it just a thing you will always get with single-winner elections?
When we consider what number of candidates is ideal in a given race, cognitive load is a huge factor. Having a candidate set larger than say 5-7 is bad in terms of voters being able to do a good job, getting overwhelmed, etc. On the other hand we obviously want to have enough good options to inspire turnout and get a representative winner or winner set.
Having a candidate set larger than say 5-7 is bad in terms of voters being able to do a good job, getting overwhelmed, etc
hence the role of party labels to provide mental shortcuts.
we obviously want to have enough good options to inspire turnout and get a representative winner or winner set.
right. I think all available evidence seems to point to the conclusion that this won't happen with single member districts.
@andy-dienes I have my preferences for things like ideal candidate set size, partisan vs nonpartisan, and PR v single winner v hybrid, as well as my preferences for ideal mechanisms to dial those all in, but that's somewhat irrelevant.
A very important design consideration for a good voting method, imo, is adaptability and compatibility. You may strongly prefer multi-winner party list PR based systems, but unless you are confident that you can get that passed everywhere all at once with no transition period, that's not a winning strategy.
The fact is that some jurisdictions are single-winner and others are multi-member, some are partisan and others are non-partisan, some are competitive and have problems with vote-splitting, others struggle with unopposed seats. Some have primaries and some have top two runoffs, some newer proposals have neither.
Some jurisdictions like some of the above system characteristics and others want some of them changed, but I think it's very rare to find a place that is open to considering a total overhaul of their entire system all at once. If that does happen (like in my hometown, Portland, OR,) the likelihood is that there will be a lot of practical and political ramifications from bring unpopular changes along with the popular ones that carried the initiative.
To me, it's really important that our conversations allow space for our voting methods and systems to slot into a variety of other electoral systems without magnifying problems, creating new problems, or backfiring and branding voting reform negatively.
Back to your original post and main point, my theory is that IRV enjoys a swell of candidates at first because it seems like it would help underdog candidates, but the reality is that vote-splitting can still happen if there are three way competitive races, which is the exact scenario candidates like that are hoping for. In competitive races they can still be spoilers, and maybe more apparent to the less savvy, is the fact that if they don't make it past the first rounds then the down ballot support for them may never be visible to the public. (see image below)
All of that is likely to leave a once starry-eyed prospective politico feeling disillusioned and not inspired to run again. Campaigns are hard and once people learn that they aren't viable most are likely to stop trying after a couple cycles.
Sara: "we obviously want to have enough good options to inspire turnout and get a representative winner or winner set."
Andy: "right. I think all available evidence seems to point to the conclusion that this won't happen with single member districts."
Based on what? Competitive single-winner elections happen all the time. Again, Portland is a good example. City elections are Nonpartisan Plurality + Top Two. The last election I voted in there had over 20 candidates for the mayor, and over 20 for one or two of the city council seats. These offices all cost a lot to run for and the jobs pay a lot as well. Partisan primaries (which are single-winner) also often have a lot of candidates, especially if the seat is vacant or if the incumbent is unpopular.
In a Top Two, the danger of being branded a spoiler is removed, so you see a lot more people willing to risk a run than they would otherwise.
Again, if you wanted to say something like 'multi-winner elections are a surefire way to increase the number of candidates per race' then we'd all be onboard, but framing it like 'my way is the only possible way' isn't compelling, and lacks perspective.
PS: I saw you edited your comment above and the new wording is a lot better, but (mod note) if we make substantive changes after a comment has been responded to, it's a good practice to note that it's been edited. Like so works well: [Edited] That way the rest of the thread still makes sense to other readers.
@sarawolk Nothing is absolute, but the fact is that over 90% of districts are completely uncompetitive. It's great that nearly 10% are (and some municipal elections are competitive---glad to hear that Portland is one of them) but that's not a particularly compelling counterpoint to me.
Moon Duchin and Lee Drutman have both written great analyses of the harm that SMD do at a structural level, both with respect to "malicious" gerrymandering and to when fair districts simply cannot possibly be drawn (famously: the Massachusetts problem). Here's Drutman's article. Here's a recent one by Duchin
I did not edit any comments except for maybe seconds after posting. not sure what you're referring to.
@andy-dienes I think we've talked enough that you know I'm not advocating for just leaving the system as it is. I can think of a host of things that could change to address that issue, including STAR Voting, so what I don't get is why you are only willing to look at that one possible solution? Text is hard for reading tone, but that's the vibe I'm getting from your replies.
Many of those uncompetitive seats are billed as "part time" gigs where you get a stipend not a paycheck. Many use the worst kind of plurality systems where vote-splitting clearly creates another strong incentive to not run. Many are rural places without that many fish in the pond to choose from. I do think multi-member districts are a great option for those kinds of seats and I've recommended them to rural commissions when it seems like a good option they'd be open to. For places where gerrymandering is an issue and it isn't solvable more directly, that's another strong argument. We agree on that.
I just think it's problematic and counterproductive to assert, or imply, that single-winner districts are inherently evil and that proportional representation is the one and only savior. Which is not what you said, obviously, but it's how these debates in general tend to come across. If that's not how you mean it to sound maybe we should rewind to some common ground?
I also want to get back to my main point above, which is that I think we should be looking for a system to encourage the number of candidates to be in the ideal range, not too few, not too many.
Ps: Nevermind about the edit. The sentence I was referring to was actually one post higher than where I was looking.
so what I don't get is why you are only willing to look at that one possible solution?
Because I think the country is running out of time to fix things. Bikeshedding about which single-winner rule might or might not be the best is a waste. Just look at what's happened in Wisconsin, where the state legislature has a supermajority with a minority of votes.
STAR won't fix that. At least, not without some very optimistic second-order effects, and even then not anytime soon.
The only path I can see that has a more than 0% chance of getting voters some actual representation within the next few election cycles is exactly what Drutman, Santucci, Shugart, and every other political scholar on this open letter is calling for, which is party-list PR. It's been my experience historically that EVC is rather hostile to list PR, so it's going to be somewhat hard to get on the same page.
like you pointed out, my comments apply to legislative elections only. for municipal elections I agree that it might be silly to use a list-PR rule. I think my existential dread is more strongly derived from the legislative elections I see than the municipal ones though so they are definitely more top-of-mind
@andy-dienes I'm less interested in bike-shedding about which one reform (single-winner or multi-member) is best, and more interested in using the best tools in the bike shed to fix the actual broken parts to get those bikes on the road as simply as possible, while setting up those bikes to be primed and ready for more upgrades down the line. What tools we'll need depends on what's broken and what the customers want. What new parts we put on the bike also depends on the type of bike they already have.
An example: My bike chain broke this week so I took it into the shop and asked for a new chain and also a new sprocket, since normally they need to be replaced together and it's an older bike. My friend working there agreed that they both probably needed replacing, and found a few other things that he'd love to fix while he was at it, but I shared that I needed the bike ASAP and that money was tight this month. We agreed on just the two things I'd asked for for now. Then today he called to say the bike was done. He'd planned on replacing the sprocket but it turned out that the new chain actually fit better than he expected and it rode great on the test ride so it's going to be a simpler fix this time. I can always do more repairs later if desired.
In some cases a total overhaul is what's needed. My housemate also took his bike in to the same shop with a broken chain, but they did $450 worth of repairs. His bike needed it. I see PR as a total overhaul for most jurisdictions. I also see it as a big piece of the puzzle in my dream total overhaul, but I'm not going to tell people that unless they are going to do that they're wasting their time on Approval or STAR Voting. That would do the movement a huge disservice.
[Edited to add] Cheat sheet: Bike chain = ballot
PS. My existential dread is pretty evenly distributed across world politics at every level.
Andy: "EVC is rather hostile to list PR"
Our board has diverse views on this, and there is room for diverse views on this topic in our movement. Please don't assume that views held by one of us are agreed on by the majority of us, or that a good argument that addressed our concerns couldn't change our minds. I'll try my best to speak for the group here:
The bottom line is that we appreciate the differences between bad systems and better systems. We just have a high bar for what reforms we endorse and what characteristics those reforms have to have. For example, reforms we support basically have to eliminate vote-splitting. Decent PR proposals that are married to bad single-winner systems are a problem. We would be very supportive of most any decent variation of Approval List PR, Condorcet List PR, or List STAR-PR, especially for places that already have a partisan parliamentary or PR type system, or those that want one.
Top-2 and List PR are both examples of what I would consider stepping stone reforms for single and multi-winner. Both get statistically much more representative outcomes with a simple ballot, and both are a relatively minimal change from the voter's perspective. Both are good stepping stones for more comprehensive upgrades down the line. Neither are what we would propose because we know voting reform is hard and we might as well get it right the first time, but I personally am more open to either of these reforms than IRV (which was not a conclusion I came to easily) and I think they both deserve a place in the movement. There is definitely room for List PR advocates in our coalition and on our team.
Back to your original post and main point, my theory is that IRV enjoys a swell of candidates at first because it seems like it would help underdog candidates, but the reality is that vote-splitting can still happen if there are three way competitive races, which is the exact scenario candidates like that are hoping for. In competitive races they can still be spoilers, and maybe more apparent to the less savvy, is the fact that if they don't make it past the first rounds then the down ballot support for them may never be visible to the public.
I don't have all the data, but my understanding is that the reason Burlington 2009 keeps coming up in discussions (despite being many years ago) is that it was just about the only election where IRV didn't elect the Condorcet winner. This suggests to me that any sort of vote-splitting hasn't been the reason why IRV doesn't increase the number of candidates.
It could simply be that people are exicted by the new method at the start, but then they see that changing to another single-winner method doesn't really change that much, so the interest goes down again.
But I agree with you that this doesn't mean that discussions should only ever be about PR. There are many different types of elections - some of which are just for a single position that aren't part of a wider body. I see this as a place to discuss the best voting method for any type of election, whether for a country's national parliament or for how a group of friends should decide where to go to eat.
any decent variation of Approval List PR, Condorcet List PR, or List STAR-PR,
I'm sure there is an interesting experiment to be had here---the optimist in me would especially love to see approval list---but the path to adoption at any real scale is a decade out or longer at best
regular, vanilla, list-PR is already legal, simple, and is incredibly well-studied. it takes almost no changes to existing law or election infrastructure to implement.
You say you are not hostile to it, but you imply otherwise with the statements
Decent PR proposals that are married to bad single-winner systems are a problem.
stepping stone reforms
might as well get it right the first time,
It just feels a little silly to me to assume that anyone here knows better than any of the signatories of this letter (and certainly not collectively!)
I'm not suggesting you should stop working on STAR advocacy: that's your thing, and I know it. I do wish that when it comes to legislative elections, EVC took a bit more of a proactive stance in working with and endorsing groups like Fix Our House in their advocacy of PR. I also hope you can understand why I might not get too excited about any kind of single-winner reform when it won't move the needle in 90% of districts. As a @Sass -ism: "it's the drawing of the districts, not the math of the method"
Sass last edited by
@andy-dienes Considering that the list of signatories is almost exclusively academics who have spent zero time in the advocacy space beyond signing a letter to a body that will ignore it, I'm not inclined to believe that they understand how reform works in the US. I truly do write off that entire letter because the people who've signed it (except for maybe Lee Drutman) have done almost no actual work to improve elections in the US. We do know better than most of those signatories because we collect signatures and create educational materials and build coalitions and read constitutions and election codes and support volunteers and so much more.
And your -ism misses my most important point: don't blame voters. It is the math of the method, not a vice of the voters or a crime of the candidates. I understand perfectly that "drawing better districts" doesn't solve gerrymandering, but I recognize that our most pressing problem is polarization, which is a direct result of the center-squeeze effect.
We advocate for STAR, Approval, Ranked Robin, etc. because we know that there's a way to eliminate center-squeeze and that single-winner reform is the fastest way there. Americans don't know anything about voting science and the laws and constitutions in place are major obstacles to ProRep. Focusing on single-winner is a strategic decision that applies to the US right now. Hopefully in 20 years, it won't be, but for now it is.
We do know better than most of those signatories
I don't have a better way to describe this than "comically arrogant." These are people who have devoted their lives to the study (and pedagogy!) of sociology and history with a lens for politics, elections, and governance. You think they don't "read constitutions and election codes" or create "educational materials"?
Let's just take one at random. I did not cherry pick anyone impressive, literally just scrolled through the list and threw a dart, and it landed on Dan Slater.
- BA and MA in IR & History from UW-Madison and PhD in polisci from Emory
- a Fulbright scholarship spent studying and working in Malaysia
- spent the last 24 years professionally studying emerging democracies and patterns of contentious politics writing 2 books and god knows how many papers
- done consultancy work with policy groups such as AEI, CEIP, OECD, Freedom House, World Bank (aka "actual work", as you put it)
I don't know exactly how old you are, but I'm willing to bet he's been paid to study comparative politics since the time you and I were in diapers.
And this is just a singular, randomly picked, scholar on the list.
When you next encounter someone who might not be taking EVC or its claims seriously, and you get frustrated or question why you're receiving that response, I think you should remember this exchange, because it's this exact attitude that will turn people away.
Saying something like "yeah, 200 of the world's top experts in this domain who have been studying this for decades all agree on
X, but I know better than them because I'm starting an initiative in local politics for
Y," doesn't make me doubt
Xit makes me doubt you.