[EL] RCV in San Francisco

Sat Jun 30 07:04:01 PDT 2018

Rob,

Thanks for correcting my calculation! I would stop raising this point repeatedly if RCV proponents stopped continuing to claim that “the goal of any runoff system is to ensure that the winner has a majority (50% + 1) of the vote and is the candidate preferred by the most voters. [Jurisdiction’s name] ‘instant runoff’ elections fulfill both goals.” When you say Fair Vote accepts the realities of imperfection, that sounds like you agree that RCV does not necessarily fulfill either goal. It would be great if advocates stopped making this argument.

I completely agree with you that RCV can be an improvement in many cases. San Francisco’s previous system of holding the runoff in December when turnout was abysmal was definitely one case where RCV is almost certainly is better. Primaries with runoffs are another. I also think adding RCV to a top-two primary system would be desirable if one wants to have a top-two primary system, for the reason you point out.

However, I think the case is less clear for other systems -- for example, when local jurisdictions have nonpartisan primaries in June and runoff elections on-cycle in November. In this scenario, the winner of the November run-off may end up with more votes than would be the case under an alternative system with no June primary and just an RCV election in November with considerable ballot exhaustion. (Whether the two candidates that make it to the final RCV redistribution round are “better” or different than the candidates who make it to the November runoff is an empirical question.)

I’m particularly skeptical of adding RCV to presidential general elections (an argument Ned Foley makes in the paper I linked to earlier), given how this would change the incentives for third-party and independent candidate entry; I can imagine scenarios where under RCV, a presidential candidate could win a state’s electoral votes with a smaller share of the total valid votes cast (given sufficiently high levels of exhaustion) than would be the case under the current plurality system used in most states.

Vlad

[The Ohio State University]
Vladimir Kogan, Associate Professor
Department of Political Science
2004 Derby Hall | 154 N. Oval Mall, Columbus, OH 43210-1373
510/415-4074 Mobile
614/292-9498 Office
614/292-1146 Fax
http://u.osu.edu/kogan.18/
kogan.18 at osu.edu<mailto:kogan.18 at osu.edu>

From: Rob Richie [mailto:rr at fairvote.org]
Sent: Saturday, June 30, 2018 9:26 AM
To: Kogan, Vladimir
Cc: Rick Hasen; Election Law Listserv
Subject: Re: [EL] RCV in San Francisco

Vlad,

I find your "broken record" repetition on this  topic overly selective. I'll make three points that include a correction and what I trust are several useful numbers

*  You suggested last week that FairVote has given up on the goal of representative outcomes -- that is, fair representation of both those in the majority and the minority. That's not true. We just accept the realities of imperfection. You set up ranked choice voting in contrast to a theoretical model of majoritarian perfection that our current systems don't remotely achieve.  Note, however, one could achieve your goal if we followed Australia's model of mandating voting and mandating rankings. Every RCV winner in Australia's House elections must earn a majority of votes from registered voters who didn't want to risk being fined and who cast a valid ballot.

* But we don't mandate voting in the USA and that's not FairVote's fight. Americans don't include the huge majorities of eligible and registered voters who don't vote in our primaries  (and most of our other elections)  in the "denominator" when we report percentages for winners. We don't report undervotes when reporting winning majorities. If we did, we'd have lots of low single digit "winners." And to be clear, the primary runoff elections that you seem to give a free pass make it WORSE on average, not better. We're now just over halfway through the primary season, and there have been 23 primary runoff elections. Notably

-- The median runoff primary winner earned LESS than their vote total in the first round. This obviously can never happen with ranked choice voting

--- The 23 winners all earned Vlad-defined "majorities" in the runoff, but all but nine of them earned less than a third of the first round primary vote. Only three surpassed the 46.2% earned by London Breed in San Francisco despite the ballot that that limits rankings to three. (Vlad is using the wrong denominator in San Francisco, as he counts first round undervotes that no one uses when determining majorities in elections that might go to runoffs.)

-- Of those three relatively strong runoff showings, only two had majorities of the 1st round: one got 50.6% of the 1st round and the other 55.4%

-- In addition the runoff candidates might not be the most representative candidates. The top two finishers in runoffs can include candidates who get to the runoff only due to split votes. We've in fact had important runoffs recently under "Top Two primary" rules where no candidates from the party with the most primary votes made the runoff - that happened in 2016 in the Washington State treasurer race (affecting Democrats) and a Georgia state senate special election (affecting Republicans.) RCV is far more reliable to have the strongest two candidates make the final round -- not with 100% perfection, but clearly an improvement over runoffs without an RCV ballot.

--- Of course, most of our primaries this year have been held with plurality rules. 10 winners of US House and US Senate primaries earned less than 30%, and another 16 won less than 40%. Two (WV-3 and PA-13) Republican primaries were won with less than 24% in heavily GOP district that Trump won by more than 50%. Interesting to think about from a view of "majority rule."

* So let's look at RCV, which absolutely outperforms plurality and runoff elections -- it's not even close. Maine used RCV for the 1st time, with very limited state-financed voter education.  Democrats had more valid votes for governor in a primary than ever before in state history. Two primaries requiring instant runoffs - the governor's race and the 2nd congressional district. The impact of RCV were:

- Governor:  Janet Mills went 33.09% in the first round to 50.2% of the first round vote total (and 54.1% of active votes in final round).

- CD-2: Jared Golden went 46.4% of the first round from an instant runoff found of 54.3% of the first round vote total - that is, a higher winning percentage than all but one of the 23 congressional primary runoff winners this year using the same measure.

- But that's not all. These candidates were also seeking support from backers of their final round opponent in a very real, meaningful way that they knew might help them win in the campaign. London Breed was ranked 2nd or 3rd by nearly half of her final round opponent's backers, giving her a top three ranking from 63% of San Francisco voters In the governor's race in Maine, Janet Mills was the top 2nd choice of backers of her final round opponent Adam Cote, and 47% of his supporters ranked her 2nd or 3rd. In other words, if you add the votes she had in the final round with Cote votes ranking here highly she had the active backing of well over 70% of Democrats. That's an impressive improvement from having only 33% of 1st choices , and a nice reflection of the value of  Mills "seeking consensus" in her RCV campaign.

So Vlad, please excuse me if I don't see your argument as a reason not to improve our elections with ranked choice voting

Thanks for reading,
Rob

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20180630/f0a65629/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 3605 bytes
Desc: image001.png
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20180630/f0a65629/attachment.png>