Some snow day thoughts on election reform since I can’t take a break from thinking about civics. The California jungle primary is starting to look pretty stupid. With all due respect to Pablo, a strong and smart advocate of that reform on this site, it is starting to create a lot of unintended consequences. First, it introduces runoffs and non-partisan primaries which are both great ideas, but in maintaining first past the post, it actually undermines its own intentions.
There is a strong chance the crowded Democratic fields in some contested California Congressional races, a sign of excitement and engagement on that side of the aisle, could actually result in Republicans advancing to the final two spots and shutting out the Democrats entirely, since their vote is more concentrated to fewer candidates. For example, in one district where there are 13 Democrats running, even if they combine to get 60% of the vote overall, the 40% of the voters split between the two Republicans would have a far higher chance of advancing to the final two spots. Thus, a Democratic leaning district that produces a stronger and more crowded primary field will be stuck choosing between candidates a minority of the electorate would have chosen. Converting these jungle primaries to a ranked choice instant runoff model would solve all of these problems, and avoid the need and expense of multiple elections.
If done in a non-partisan way, IRV would also allow us to pick and choose candidates from either party that we like, which should help prevent the emergence of extremists candidates in the future. In the Illinois Attorney General primary there are five independent Democrats running who will likely dilute the vote on that side of the fence letting one of the hackish establishment Democrats win with a minority plurality instead. There is also an interesting socially moderate, biracial Republican running on a strong criminal justice reform and ethics accountability platform who I would absolutely want to vote for as my sixth choice in a hypothetical open primary IRV ballot. Thus, if none of the five Democrats I liked advance to the final, my vote would transfer to her.
Without getting too deep in the weeds, it is far less likely the GOP would have nominated Trump in an internal primary with IRV, let alone, an open primary where Democrats could vote for Kasich as their third preference after Bernie and Hillary. Not to mention all those Stein and Johnson votes would have transferred to Hillary instead of being ‘wasted’ on ‘spoilers’. As the California example is showing us, along with the worldwide populist revolt against the establishment, it may start to be in the establishments own interests to open up election systems and allow for more cooperation, consensus, and true majority rule rather than undemocratic closed primaries and the spoiler effect marginalizing outside of the box candidacies.
Even making partisan races convert to IRV would have a far better effect on our system of government than a jungle primary. In Massachusetts the same dynamic is playing out in our contested Congressional races, which also currently limit independent and third party participation by forcing independents to pick a partisan ballot and keeping third party members out of the process entirely. It is likely the winner of the CD-3 race will be selected with a plurality of votes so low that a supermajority of the district selected somebody else. Ranked choice would also do a lot to keep the Capuano and Pressley race civil, since essentially you could vote for both candidates and just pick the one you like better to be your first overall. It would also keep it cheap, since the candidates would lose any incentive to run attack ads funded by outside interest groups.
Ranked choice would also help in our primary for Governor, since we could vote for all three candidates in the order we like the best. Thus my vote for Berwick last cycle would not have ‘spoiled’ a second choice preference for Grossman. A vote for Massie this cycle would not ‘spoil’ a second choice preference for Warren and even by forcing me to vote for Gonzalez third, if he ultimately wins I would know I could live with him (after all I voted for him too, even if I voted for him last) and that he had a true majority of the party behind him.
Ranked choice voting is awesome and about to happen in Maine where we can see it take place. I hope we can pass it here.
Excellent analysis, James. I, too, think top-two was a worthwhile experiment, but one that just hasn’t worked out that well.
In addition to the problems its causing for the Democratic Party, there is the lower-case “democratic” problem that it excludes voices and choices from the general election. This has mostly hurt third party candidates, but it’s beginning to hurt major parties too, as you point out. It may also be a key reason behind California’s declining voter turnout. With Ranked Choice Voting (aka IRV), we can have our cake and eat it too — more voices and choices but no spoilers.
Voter Choice Massachusetts is bringing Ranked Choice Voting to Massachusetts. Everyone who wants to see this happen should sign up!
I am glad that both you and JConway now agree that ranked voting is better than jungle primaries. But the results of this “experiment” were both predictable and predicted.
I’ve said this before when this came up in the past, but my main concern with ranked choice is that it will make voting harder and therefore reduce participation. Case in point is Cambridge local elections, which many people I know skip because it’s too hard to rank 30 people. A possible compromise would be to have a ranked choice primary and then a general election between the top two.
That is a curious thing to say. Turnout for Cambridge city elections has been consistently high (in recent years, more than a third of all registerd voters). This despite the fact that a “weak mayor” system like Cambridge is thought to attract fewer voters.
I’m not sure why Cambridge does so well. Perhaps it is a voting system in which every vote elects someone.
Fair enough. I guess what my post should say is “people who wouldn’t vote in municipal elections anyway say they don’t vote in such elections because it’s too confusing, if they live in Cambridge.”
It’s a legitimate concern, but as Trickle up said, Cambridge actually has higher local turnout than other cities. They also have very low ballot error rates, comparable to plurality elections. Ranked choice has also delivered very diverse representation to their city council — very unlike, say, Lowell, where people have color have largely been shut out of the City Council due to a lack of RCV.
Yes, the Cambridge ballot is more complex, but that’s in large part because they’re electing 9 at a time, and they allow you to rank all the candidates. All state and federal offices in Massachusetts elect only one (and most multi-winner applications around the world usually call for electing 3-5 at a time). In sum, the ballot complexity in Cambridge is the very upper limit of RCV complexity, much higher than what you would see if we used it for state and federal offices; and from the data Cambridge voters seem to be handling even that level of complexity well.
More details here.
Bingo. I am specifically talking about an instant runoff system for single member districts in legislative races or executive positions.
It is important to note that Cambridge is also not strictly speaking the same ranked choice system we are talking about here. Proportional representation via single transferable vote for 9 at large seats works out closer to a form of fractional voting than the kind of instant runoff we are talking about for single member districts and executive positions.
I used to be a critic of the PR-STV voting method, but it has resulted in a more diverse city council body. The recent re-emergence of slates has also helped contribute to greater civic participation and more cooperative campaigning as the 1940 adopters of Plan E and PR had intended.
I agree with Trickle Up that switching to a Plan B system with a strong mayor and ward representation would likely be a better model of municipal government for a city that size and those seats and the mayors race could still be conducted via IRV.
As a former Cantabrigian, I actually think that Cambridge’s city charter (weak mayor, strong manager) is just fine. The form of IRV that it uses obviates the need for wards, too.
My comment about that was that there is evidence that strong-mayor races excite greater interest, all else equal, than elections for “mere” city councilors.
Consequently Cambridge’s relatively high turnout compared to that in comparable cities with strong-mayor charters suggests that their unusual method of voting does not discourage turnout and may even encourage it.
In Somerville, turnout is usually driven by the Ward races. If there’s a contested race for Ward Alderman, turnout is high in that Ward, if there isn’t, then turn out is low, often abysmally low. Last fall was something of an exception, when we had decent turnout even in the Wards without contested races.
If the two entrenched parties hadn’t conspired to limit people’s information in the last presidential election by closing the televised debates you might be celebrating Hillary’s presidency now. There is a lot of mistrust of the system and IRV will only make it worse.
You think voters for Gary would have picked Hillary as their next choice?
If you’re on a ballot debates should not be able to exclude you.
Strong agreement about the ridiculousness of the presidential debates. If you’re on enough state ballots to have a mathematical chance of winning, you should be included. (I may be open to a small polling threshold if the number gets unwieldy, but we already tolerate presidential primary debates with far greater number of candidates).
One of the things I really like about IRV is that it flips the debate incentive. Under plurality, Clinton wants Stein excluded; under IRV, Clinton wants Stein included. If Clinton thinks she would be the second choice of most Stein voters, then she wants as many Stein voters to turn out as possible, and she will be respectful towards Stein to win those seconds choices. What’s in it for Stein: a chance for her ideas to gain traction and a chance for her party to earn more support and grow.
Greg put it really well. I think IRV is actually something the major parties should consider supporting. Especially the Democratic Party which should support this reform in principle since it expands democracy while also supporting it pragmatically since it has arguably been hurt more by the third party in recent years.
Well to be a little cynical, though I would argue realistic: Ranked voting does remove barriers to third parties. For that reason it threatens the duopoly.
It’s not irrational for the two established parties to resist that. In fact, I expect it.
If IRV encourages the Jill Steins of the world, it’s a bad thing. The beauty of the Top Two primary system is that it removes the fringe candidates before we hold the final election, and we can focus on the two strongest candidates.
The beauty of IRV, on the other hand, is that it removes minor parties as spoilers, and lets people articulate their actual interests through the voting process safely and without lesser-evil trade-offs.
Top Two on the other hand has the significant possibility of selecting two finalists whose combined first-round vote is less than 50%. It can boost extremists, and it poses unplayable game-theory choices to the voters.
Arguably the National Front has only been enhanced by sneaking into the final round on multiple occassions, whereas in an IRV runoff it would never have made it that far. There are numerous examples of runoffs helping fringe candidates in other parts of the country and there is a real risk California Democrats will be punished for attracting a wider and more compelling field.
That is not the case in California where an excited and mobilized Democratic primary field will dilute the overall Democratic vote and allow two Republicans to advance to the final in Democratic leaning or fair fight districts where Democrats should at least have an even chances
IRV would also allow for third party candidates to take protest votes and have those votes redistributed to more viable candidates in the same election. In a jungle primary that would only happen after the first round and there is no guarantee both major parties would advance to the final.
So adopting RCV in presidential races would actually incentivize greater third party participation since it eliminates most aspects of the spoiler effect. The electoral college still poses a challenge, but this reform would actually allow voters in swing states an opportunity to vote for third parties without “throwing their vote away”.
The debates are a separate topic since a private organization controlled by the two major parties is a fairly undemocratic way to run debates. I supported allowing third party candidates to debate in 2016 and believe an RCV model would also increase their rate of support and make for more civil and cooperative debates. Much as all the party candidates debated in France for the first round, they could do the same under a RCV systems.
Every Johnson supporter I know voted for him only if they lived in a safe state, they voted for Hillary if they lived in a swing state. This along with exit polling indicates that she was he likelier second choice. Then again, that is not the primary reason I would cite to adopt a more democratic voting method. It would help third parties enormously, and the MA libertarians I know like Dan Fishman are working with the RCV.
This is just not even a blurry semblance of reality as both Gary Johnson and Jill Stein increased their 2016 returns over their 2012 returns. Having allowed them access to the debates could only have had the affect of further increasing their returns and likely resulting in something like a Trump 45% – Clinton 47% split rather than the 46% – 48% split that happened.
Marginal is as marginal does…
We are once again wandering down the path of elevating the Evan Falchuks and Jill Steins of the world. The purpose of an election is not to elevate Falchuk or Stein, it’s to choose a candidate to serve in public office.
Let’s remember where we are. Massachusetts is not a two party state, it’s a one and a half party state, where the GOP is capable of winning only the corner office and an insignificant minority of seats in the state legislature. The biggest problem in Massachusetts is not too many choices, it’s not enough choices, with an overwhelming number of uncontested general elections across the Commonwealth.
A “jungle primary” would, in most cases, put two Democrats on the ballot. Voters would have two significant choices in November, not an uncontested or meaningless choice of one Democrat and one Green-Rainbow candidate.
IRV does the exact opposite of what you argue it would do here by forcing protest voters to select a second choice that would gain their vote in the event of a runoff. It basically does the same weeding out a jungle primary would do except in a single election rather than multiple rounds.
It does so with a major difference being that the winner of an IRV race is assured of getting 50+% of the vote and being the overall majority winner. There are still numerous examples where a plurality jungle primary results in a minority vote winner taking the election.
We have both problems. We have long streaks of uncontested elections, punctuated by hyper-competitive elections of “too many” options where a winner too often walks away without anything close to a majority. Ranked choice voting helps solve both problems. It both ensures the winner has a majority of the vote and encourages more candidates to run.
RCV encourages robust political dialogs of diverse viewpoints, conducted in a collaborative and engaging way. In contrast, top-two limits our world to at most 2 options in perpetuity, 2 options likely to battle each other with negative attacks that turn voters off; and does create this very real incentive to discourage candidates to run.
How about this? We use RCV for general elections and primaries, and then have a discussion about whether it would make sense to eliminate the primary entirely and have a single general election under RCV? I would be very open to that discussion.
I find it interesting that the biggest party is actually not a party at all. IMO people are so disgusted with the corrupt politicians we have they’re ashamed to even have a party designation next to their enrollment. We’re always voting for the best of bad choices.
In the midst of this post, you get the specious argument of:
You need to do some very creative electoral math to find a race in which there are 13 Democrats so evenly dividing the vote to the point of where one candidate won’t get more than 20% of the vote, assuming that the two Republicans evenly divide up their 40% of the ballots. I would invite anyone to find the Democratic primary where a 13 candidate field has sufficiently divided the vote in a district with sufficient GOP strength for this to be a reasonable outcome.
I could easily see that happening in CD3 if there were still a second GOP candidate. My overall beef with the jungle scenario is that I like in principle the idea of two major party nominees going head to head.
However, there are many districts where there are not two major parties. In districts with two viable parties, two candidates will emerge from different parties. From districts tilted toward one party, those voters will still have a choice in November.
Look at the link in a California and there are five or six districts the jungle primary could really dilute and hurt Democrats.
I doubt that money, volunteers, and local enthusiasm will distribute evenly among all Democrats to the point that two Republicans would advance in a Democratic leaning district.
If a Democratic district does, for some reason, have two Republicans on the ballot, the district would self-correct in the subsequent election. Either that, or one of the Republicans will move far enough to the left to gain support in the district.
It is a very serious concern, so much so that the DCCC is discouraging good candidates from running for fear the field is too wide and the narrower spread for Republicans will cause the exact problem you say is an impossibility.
In this case it also seems that the jungle primary system is punishing the party that has broader levels of support and enthusiasm. It is possible as you suggest for here that the statewide races result in two Democrats advancing past the runoff.
That happened with the US Senate race last year and could well happen with the Governors race this year. Other Congressional primaries have done so in more Democratic leaning districts, arguably amplifying the issues of wasted resources others have brought up on another thread here,
The prospect of even one of the six or seven open races in swing districts being lost because of this system strikes me as a blow to small and big d democratic principles alike. IRV allows for a cooperative rather than competitive campaign and forces major parties to compete for the votes of third party leaners and vice a versa. It would also moderate the Republican Party as they could compete to be the second choice or third choices of voters. It will never result in a scenario where a candidate wins a race with a majority of the district supporting another candidate. A feature of the first past the post system and a bug of the jungle primary.