[EM] A simpler approval based way of replacing the CA jungle primary

Rob Lanphier robla at robla.net
Fri Aug 17 14:10:22 PDT 2018

Hi Jameson,

Welcome to the party!  \o/  :-)  I've got some ideas for how to get
traction for this in California.

More inline....

On Thu, Aug 16, 2018 at 10:40 AM Jameson Quinn <jameson.quinn at gmail.com> wrote:
> I think that 3-2-1, though it's an ad hoc combination system, has something to add here.

Hmmmm....a 3-2-1 primary.  On one level, I like it.  3-2-1 makes some
cool tradeoffs between expressiveness and simplicity.

Still, in thinking about what seems tractable from an activism point
of view, it seems too complicated.  3-2-1 seems like something that
would work in San Francisco or a smaller municipality in the Bay Area,
but it seems like a big leap for a statewide replacement for top-two.

> -Initially restricting the race to the three candidates with the
> strongest top-rating support, as long as that represents at least
> two parties and as long as #3 has at least half the support of #1, is
> sensible. In practice it prevents dark horse winners, either accidental
> through "honest" but poorly-considered voting or through misfired
> strategic voting.

What mechanism would you suggest for ensuring that at least two
parties are supported in those top three?  Any why?  For example, what
would be wrong with two Democrats and an independent being the three
finalists, or for that matter, three Democrats who each build very
different coalitions and that (in sum) represent 95% of the

My Feb 28 proposal on this list had a more explicit mechanism
involving parties.  That proposal tries to use a game theory trick or
two to ensure sincere voting.  If you believe that parties should be
an explicit part of this, I'm curious if you think there's a 3-2-1
hybrid that can be combined with that proposal.

> -Of those three, the two with the fewest strong rejections should be
> finalists. This prevents center squeeze. And in practice, assuming
> the risk of backfire prevents strategy from reaching critical mass,
> it resolves the chicken dilemma by putting it off to the strategy-free
> third stage.
> -Between the two finalists, choosing the pairwise winner is the only way
> to prevent strategy from affecting this last step (considered by itself).

This is the part that I really like about this over my July 18
proposal.  I like the guarantee of exactly two candidates while also
being mindful of the center squeeze.

> The pairwise step could be a separate election, especially if neither
> candidate had a majority (or some other threshold) of ballots ranking them
> strictly higher. This could then be seen as either a "primary/general"
> system or a "election/runoff" system.

As I've been thinking about this system over the past day or so, I've
warmed up to it a bit.  I like that it would replace the primary stage
of the top two primary, while leaving the general election untouched.
This seems like something to present to an activist working group as
one of a handful of options, which would *also* include an option with
a simple approval-style ballot.  But I think it's important to ensure
a system where there would be little sympathy for a sore loser, while
also not practically guaranteeing Democrats or Republicans a spot on
the ticket.  As we learned with Burlington 2009, a good sympathy
generator is "the system is too complicated".  I worry that a 3-2-1
hybrid would be vulnerable to that charge.


More information about the Election-Methods mailing list