Ranked choice voting is going to get a high-profile test in New York City’s mayoral contest next month. Some folks are worried that New Yorkers don’t know enough about ranked choice voting to make it work, but come on: Minneapolis and St. Paul have been using it for a decade. Several other cities plus the state of Maine use ranked choice voting, and 23 Utah cities are trying it out this year. If Mormons can do it, so can Gotham!
To put ranked choice voting into a South Dakota context, some friends of the blog have put together an online exercise based on South Dakota history. The poll asks voters to pick the “best former South Dakota politician” from among these seven politicos:
- George McGovern
- Dick Kneip
- Bill Janklow
- Larry Pressler
- Tom Daschle
- Dennis Daugaard
- Stephanie Herseth Sandlin
Yeah, yeah, save your fussing about who else should’ve been on the list (George S. Mickelson! Tom Berry! Peter Norbeck! Richard F. Pettigrew!) and vote… not just for your one favorite, but for your second choice, third choice, fourth choice, and so on and watch what happens!
As of my check this morning, there have been 72 votes. On the first count, McGovern, Janklow, and Daschle come out close at the top as first choices, but no one breaks 30%, let alone the 50% required to win.
Thus, by RCV rules, we drop the lowest vote-getter (Pressler) and anyone who got no votes (Kneip), and count again. In Round 2, the ballots of the Pressler pickers now go to those voters’ second choices… and that’s where we start seeing the interesting preferences:
The Pressler voters appear to lean Democratic. Absent Pressler, they turn to McGovern and Herseth Sandlin in apparently equal numbers. Assigning their second choices keeps McGovern at the top, but they aren’t nearly enough to bring McGovern over 50%. Nor are the Pressler second choices enough to raise Herseth Sandlin above Daugaard. In Round 2, SHS is the new last placer, so now she drops from contention and we count again! Whee!
SHS’s voters now see their votes go to their second choices. Pressler people who picked SHS second now see their third choices count. That collection of preferences boost McGovern, Daugaard, and Daschle. But frontrunner McGovern still is nowhere near 50%, so we drop new last-placer Daugaard and go to Round 4:
Finally, the Daugaard voters give Janklow some second-place action. Janklow gets the biggest bump now, as we’d expect from the elimination of the most recent and most faithful Republican on the list. McGovern creeps up again, but dang! the top three remain as tight as they were in Round 1, no one has a majority, and the remaining bottom two are tied. What do we do?
This RCV exercise resorts to the same democracy-thwarting method we use in current election ties: we flip a coin. Janklow wins this toss, so Daschle is out, and his second choices (and likely some Daugaarder third choices, SHSer fourth choices, and maybe even some Presslerer fifth choices) now count in the final battle:
Dropping Daschle pours a whole lot of second-bests into McGovern’s pile and lifts him to a 61% majority.
This exercise demonstrates that candidates would want to approach a ranked choice voting contest similarly to how they approach South Dakota House races, where candidates compete for two available seats. In such a race, you don’t necessarily have to be everyone’s favorite, but you want to be likable enough to be lots of people’s second choice. In this exercise, eventual winner McGovern ranked first among only 29% of voters, but a lot of his opponents’ voters were willing to settle for McGovern as second best. In contrast, Janklow had nearly as many first-place votes as McGovern, but he didn’t win much second-choice support from his opponents’ voters. This small and unscientific sample thus seems to affirm the conventional wisdom that voters either loved Janklow or hated him and suggests that McGovern was a more likable guy.
Notice one factor that didn’t pop up until the last vote: “inactive votes”. It appears that three of the 72 participants in this poll got tired after four votes and didn’t mark a fifth choice. They marked Pressler, SHS, Daugaard, and Daschle in some order, then looked at the remaining names on the ballot and didn’t find the enthusiasm to click one more box. If voters don’t mark their rankings all the way down, it is possible that a ranked choice vote requiring more rounds to reach 50% will eliminate undermarked ballots from the final count. So conceivably, a candidate who wins by ranked choice voting may not win by a majority of all ballots cast, but only by a majority of all ballots cast by people who cared enough to mark their preferences all the way down the ballot.
But apathy whittles the majority threshold in regular elections as well. So under ranked choice voting, if you really don’t like a candidate, even if you don’t have a first choice, you’d better mark your “absolutely under no circumstances” turkey last and then fill in numbers for everyone else on the ballot.
But hey, click on the survey, try out ranked choice voting, and see how much you can skew this fun exercise in picking South Dakota’s former politician.