I have a new report with Kevin Kosar and Jaehun Lee of the American Enterprise Institute:
Key Points
Portland, Oregon, used its decennial charter revision process to overhaul its method for electing the city’s legislators in 2022.
The new system divided the city into four districts, with each district represented by three members selected through ranked ballots.
Portland first used this new system in November 2024, and initial analysis indicates that it expanded the representativeness of the candidate pool and resultant council.
Here is a 2023 blog post on a “downward cascade” process in that year’s election results.
I wrote about this system in a 2021 article, which also covered its historical use in Australia and similar rules at other times in the U.S.
McCune et al. (2024) have an interesting article using the block-preferential (or “sequential”) procedure to show how it would change observed results under Scottish STV.
One alternative to outright repeal would be to replace the system with a “bottom-up” RCV allocation. The issue in Utah has been multiseat districts, which formerly were combined with runoffs.
An emergingnarrative holds that ranked-choice voting is in trouble as a cause due to its performance at this year’s general election. It therefore might be useful to look at historical data on the incidence and potential consequences of ‘bad years.’ I don’t see much evidence for the effect of a ‘bad year’ — at least from the perspective of failed adoption.
The data cover efforts to adopt the single transferable vote (STV), in almost all cases alongside the council-manager form of government.1 Close readers will note that this is not the same as what lost earlier this month: instant runoff voting with jungle primaries. I do not have the kind of data you see above for “single-winner” adoptions (i.e., instant runoff), at either the state or local levels. What I can say is:
I am not aware of any effort in this period to impose nonpartisan instant runoff for statewide elections.2
Statewide-election use of instant runoff was restricted to party primaries. I do not know if this was in state law or a decision internal to parties themselves.
I think these data provide a decent if imperfect comparison because we are dealing with the same basic phenomenon: an effort to break up parties so that more independents might win.
The data show a string of bad years in the late 1930s. They clearly did not end adoptions; more than half of winning referendums came after. Also, these ‘bad’ years followed a few in which STV did very well.3
