[Election-Methods] Dopp: 2. “Requires centr alized vote counting procedures at the state-lev el"

[Abd ul-Rahman Lomax]

At 08:11 PM 6/12/2008, Jonathan Lundell wrote:
>On Jun 12, 2008, at 12:53 PM, Abd ul-Rahman Lomax wrote:
>
>>But it isn't so fast, necessarily. In San Francisco, one election
>>required 19 rounds of eliminations, as I recall. They took a month,
>>I think, to issue the results.
>
>SF has always been notoriously slow to finish their counts and certify
>the results. More than one registrar has lost their job over it. Do
>you have reason to believe that it's actually taking significantly
>longer since IRV was implemented? The 1999 city election took three
>weeks to count.

Yes. See this: http://www.sfgov.org/site/elections_page.asp?id=68997

Common practice is to hand-count the votes in each precinct using 
election officials there. But because of the RCV requirements, I'd 
guess, they took all the ballots to City Hall to count. I'm *not* an 
expert on this election, to truly research it would require much more 
digging than I've done.

In the previous RCV elections, before Nov. 2007, a report was issued 
regarding the vote transfers. To my knowledge, that report is still 
missing. From the Excel Statement of Votes, though, one thing is 
going on. There has been a drastic increase in the number of 
candidates, it seems. There were twelve on the ballot in the Mayor's 
race, plus six write-ins. (San Francisco doesn't allow unregistered 
write-in candidates, but some do register as such.) The Mayor's race, 
however, was won in the first round by a large majority.

Turnout in Nov. 2007 was pretty low, 35.62% of registered voters. San 
Francisco sometimes sees twice that, I think.

Continuing to look at the Mayor's race, Newsome got 105,596 
first-choice votes, out of a total of 149,465 ballots cast, with 
5,627 undervotes and 479 overvotes. The runner-up in the first round 
got 9076 votes. One of the write-ins got one vote in the first round. 
Embarrassing. None of the six write-ins got more than nine votes.

There were three elections with RCV on the ballot: the Mayor, the 
District Attorney, and the Sheriff. Kamala D. Harris was unopposed on 
the ballot for District attorney, but there were write-in votes 
recorded. Harris got 114561 votes; there were 33158 undervotes, 
people who didn't vote for the office at all. All the write-in 
candidates together got 1744 votes. (All these are first choice 
votes). A landslide.

For Sheriff, Michael Hennessey got 95948 votes to David Wong's 34031 
votes, and there were 221 votes for a write-in, and 362 overvotes and 
18918 undervotes, in the first round. A landslide.

No instant runoff rounds were necessary. So why was this election so 
difficult to count? Well, obviously, they had come to depend on the 
OpScan machines. They didn't have the staff arranged to do rapid 
manual counting. Where manual counting is used, precincts generally 
count the votes the same day as the election, into the night, most of 
them. Many hands make small work.

But come the 2008 election, we'll see what happens. There will be 
Supervisorial races, and those tend to have even more candidates, and 
there have been, as I mentioned up to 19 rounds to resolve an instant runoff.

>>A far simpler method, using the same three-rank ballot as IRV, but
>>far more flexibly, would be Bucklin voting. And much, much simpler
>>to count. While FairVote claims that Later-no-harm failure for
>>Bucklin will cause wide strategic voting (bullet voting), I think
>>that actually quite unlikely. These are nonpartisan elections.
>
>While that's true in theory (all California local elections are
>nominally nonpartisan), it's far from true in fact, at least in
>political cities like San Francisco. Case in point: when now-mayor
>Newsom was running against Matt Gonzalez, and looked like a real
>threat (Matt forced a runoff in those pre-IRV days), Bill Clinton and
>Al Gore both came to town to campaign against Gonzalez, a registered
>Green. So did Jesse Jackson, Dianne Feinstein, and Nancy Pelosi.

The political races are the exceptions. Now, isn't that an 
interesting fact? Gonzalez forced a runoff. What happened? I'd look 
it up if I had time. (If the Green was in the runoff with the 
Democrat, this shows the potential of real runoffs. An opportunity 
for the Greens to show their real strength. If they had it.)

I'm a little suspicious of how this story is told. As we know, in 
most partisan races, in a big town, a Green wouldn't have a chance. 
So even if there were partisan issues here, sounds to me like this 
election was different from your standard partisan election.

>Nonpartisan? Hardly.

I'd need to see more to be able to agree with that. But that is one 
election out of many. Most of the IRV elections are for Supervisor. 
Bill Clinton doesn't typically come to town to campaign for a 
Supervisor. Are those partisan? I don't think so. And those are the 
ones going into instant runoff, mostly, because of the large numbers 
of candidates and more closely contested races.