Back to school! The question has come up in my classrooms and conversations at the ongoing APSA conference. Here are some thoughts.

1. Yes, Brad Lander and Zohran Mamdani concluded a “1-2 deal” that would not have been possible, strictly speaking, without a ranked ballot.
2. Without a ranked ballot, the Lander-Mamdani deal might have looked instead like Lander telling voters who trust him to vote for Mamdani.
3. Yes, it is possible that RCV let Mamdani stay in the race whereas, under the old primary-runoff system, an “invisible primary” would have pushed him out.
4. Regarding the “invisible primary,” it is not clear that there would have been one without RCV. The New York City party system strikes me as having been in breakdown at the time of the election. This was evident in the Democratic establishment’s struggle to coalesce around a Mamdani alternative.
5. Mamdani is part of a “wave” of Muslim-identifying and/or self-described socialist candidates. Some are winning primaries not conducted under RCV.
6. Large-N work suggests that RCV has little effect on the election prospects of “new faces.” Mamdani’s victory can be seen as consistent with this research because he already held a seat in the New York State Assembly.
7. Further evidence of the party system’s breakdown, and against the independent effect of RCV, might be the circumstances of the latter’s adoption. Why did New York City adopt RCV when it did? Why did it adopt it for nominating primaries only? Decision makers had been exposed to advocacy for many years. One possibility is that its adoption was meant to help manage anticipated coordination problems. Those problems ended up insurmountable (see, e.g., the scandals).
8. The party that came out of the election is not the same as the one that went into it. No party ever is. However, when we look back from some point in the future, I expect the historiography to emphasize elite turnover and generational change-of-guard.
9. Always remember that we cannot observe the same election without RCV and that it is important to be careful about drawing conclusions from single cases. Put simply, the risk is seeing what we want to see in the data/story/etc.