Top-k primaries might be good?
-
SNTV has two nice properties for primaries:
- It picks candidates with almost-random ideology, mimicking sortition but still giving a quality filter.
- It gives a boost to candidates with small but passionate bases of support, who see their candidate as unique and irreplaceable. This counteracts some of the mild "centrist bias" we see in strategic score and Condorcet (increasing the chances of picking the utilitarian winner).
I'd suggest 2 changes. First, reserve a spot for the incumbent. Second, assign seats to parties using a rounded-down Hare quota based on party registration—e.g. if 45% of voters are registered Republicans, and there's 5 candidates on the ballot, 2 of the places go to the first- and second-place finishers on Republican ballots.
I suggest this as an improvement on approval-with-runoff, if legislators are insistent on having two separate elections. This maintains the desirable properties of approval (unlike approval-with-runoff, which introduces turkey-raising).
-
I'd actually like to see simulations on this—ideally, I'd like a first-stage primary that filters candidates by quality but not ideological fit (which should be almost random to keep representativeness).
-
@lime said in Top-k primaries might be good?:
(regarding SNTV changes)reserve a spot for the incumbent. Second, assign seats to parties using a rounded-down Hare quota based on party registration—e.g. if 45% of voters are registered Republicans, and there's 5 candidates on the ballot, 2 of the places go to the first- and second-place finishers on Republican ballots.
I am afraid that this would turn voter registration drives into a major industry. Abuses and mistakes could occur. Misunderstandings caused by language barriers could cause noncitizens to be registered by mistake, they would not know what happened but they could be prosecuted.
Perhaps I exaggerate, but people will pay dearly for any electoral advantage.
Also, a lot of people are not aware of which party the are registered with, if any.
BTW The common term for a payment to an person who gathers voter registrations is called a bounty. So if party R is paying $10 per registration from citizens residing in district x, the bounty for those registrations is $10.
-
@gregw said in Top-k primaries might be good?:
@lime said in Top-k primaries might be good?:
(regarding SNTV changes)reserve a spot for the incumbent. Second, assign seats to parties using a rounded-down Hare quota based on party registration—e.g. if 45% of voters are registered Republicans, and there's 5 candidates on the ballot, 2 of the places go to the first- and second-place finishers on Republican ballots.
I am afraid that this would turn voter registration drives into a major industry. Abuses and mistakes could occur. Misunderstandings caused by language barriers could cause noncitizens to be registered by mistake, they would not know what happened but they could be prosecuted.
Perhaps I exaggerate, but people will pay dearly for any electoral advantage.
Also, a lot of people are not aware of which party the are registered with, if any.
BTW The common term for a payment to an person who gathers voter registrations is called a bounty. So if party R is paying $10 per registration from citizens residing in district x, the bounty for those registrations is $10.
There might be a small advantage to running more candidates (one of them might be a slightly stronger candidate than the others), but I'm guessing this is probably a small effect—generally, all the members of a party will do about equally well. There might even be a slight push in the opposite direction, because running more candidates splits funds and volunteer efforts between them. (Besides running the risk of party infighting.)
My guess is parties won't decide to do unethical things for such a small benefit.
