District 5: How Might Ranked Choice Voting Affect the Supervisor's Race?
Letís look at a real-world scenario from the upcoming November 2004 supervisorial races. Letís imagine the District 5 race with 10 candidates, four of them strong progressive candidates. What this example shows is the importance of voters using all three of their rankings, with the first choice being their favorite candidate and the second and third rankings being their runoff choices in case their favorite candidate doesnít win. Use of all the rankings is what facilitates coalition-building.
The most immediate difference between RCV and the previous December runoff is that with the old system progressive candidates likely would have had to battle it out among themselves to see which one made it to the runoff. If they split the progressive vote too badly, itís possible that none of them would have made it to the runoff. And after attacking each other, they may have had difficulty uniting the progressive vote for the December runoff.
But with RCV, the incentives for winning are different. The way one of the progressive candidates will most likely win in District 5 is by using the rankings to build coalitions among all the progressive candidates. These coalitions can be conscious and public, or tacit and subtle. A public coalition will have candidates saying to their supporters: ìLook, I think Iím the best candidate but in case I donít win, Robert or Ross or Dan also is a great candidate.î The coalition-building might extend to organizations, or even the candidates themselves, publicizing a ranking of their preferred candidates, or perhaps just a slate of three candidates with no recommended ranking, allowing voters to decide the actual ranking for themselves. Or the coalition-building can be more subtle. What the coalition candidates want to avoid at all cost are bitter negative attacks or hit pieces against each other, because it will hurt their chances of picking up second and third place rankings from the supporters of the candidates they have attacked.
Note that ìbuilding coalitionsî does not mean that the coalition candidates can not disagree with or criticize each other. It means you do it from a principled place and in a principled manner. In fact, a candidate might win the respect of voters for showcasing HOW to disagree with an opponent. RCV encourages this kind of healthy behavior on the part of candidates, and encourages debate of issues and ideas, and candidates finding common ground as much as possible.
Hereís an example for District 5, one of the most
progressive districts in San Francisco.
Imagine 100 voters and ten candidates: ProgA (Prog meaning
progressive candidate), ProgB, ProgC, ProgD; ModA (for Moderate
candidate), ModB, Conservative, MinorA, MinorB, and MinorC (Minor
candidates with no chance of winning). Letís
say after the polls close and voters have ranked their ballots, these
candidate have the following number of first-place rankings:
With 100 voters, a majority is 51 votes, so no candidate has a majority of first-place rankings, not even close. Itís a very split field. Not only that, but a moderate candidate is in first place, because the progressive vote has splintered among four candidates who, combined, have 56 votes, more than a majority. With the old December runoff system, candidates ModA and ProgA would have advanced to the December runoff, even though between them they only have 39 votes (39 percent of the total).
So now we begin the instant runoff.
Candidate MinorC has the least number of first-place rankings and
is eliminated from the runoff. In
fact, Candidates MinorA, MinorB and MinorC together only have five votes,
and because their votes do not equal the candidate vote total of the
next-highest candidate (Conservative with 8 votes), they cannot
mathematically win and so they can be eliminated at the same time from the
runoff. Of the five voters who ranked them first, three gave their
runoff/second ranking to ModA and two gave it to ProgB.
Those votes are added to the pile of the ìcontinuing
candidates,î and at the end of this round the vote totals are the
Third round: No candidate has reached the majority
threshold of 51 votes. So the candidate with the least number of total
votes, Conservative, is eliminated from the runoff. Each ballot for the
Conservative candidate now counts for the voterís next choice, which is
their runoff choice. Of the eight voters who ranked Conservative first,
six gave their runoff/second ranking to ModA and two gave it to ModB.
Supporters of the conservative candidate likely would list one or
both of the moderate candidates as a second choice, being the continuing
candidate closest to their political views. Third round vote totals are
now the following:
Fourth round: We
still donít have a majority winner, so the candidate with the fewest
votes, ProgD, is eliminated, and each of her ballots is distributed to the
voterís next/runoff choice. These
ballots would likely go to other progressive candidates.
Letís imagine that ProgB picks up 6 of the 10 votes and ProgC
gets the other 4. We now have
the following vote totals:
Fifth round: Now
the candidate with the fewest votes is ModB.
Letís imagine that ten of these ballots go to ModA, but a couple
votes go to ProgA, who is acceptable to two moderate voters because heís
good on neighborhood issues. Note that two of the votes going to ModA were
from voters who originally had ranked Conservative as their first-place
ranking, but then Conservative was eliminated and their vote went to their
runoff/second ranking, ModB, who also now has been eliminated. So now the
vote of those two voters goes to their next runoff choice, i.e. their
third ranking, who is ModA. Here are the new vote totals:
Sixth round: Still no candidate has reached the majority threshold of 51 votes. Now ProgC gets eliminated, and not surprisingly those votes go to the other progressive candidates as votersí runoff choice. Twelve voters give their runoff/next-highest ranking to ProgA, and four give it to ProgB. Note that four of the votes going to ProgA were from voters who originally had ranked ProgD as their first-place ranking, but then ProgD was eliminated and their vote went to their runoff/second ranking, ProgC, who also has been eliminated, so now the vote of those four voters goes to their next runoff choice, i.e. their third ranking, who is ProgA. Those votes are added to the pile of the continuing candidates, and at the end of this round the vote totals are the following:
Seventh round: Still
no candidate has reached the majority threshold of 51 votes. Now the
candidate with the fewest total votes is ProgB.
Of the 28 voters who voted for ProgB, most of them (26) gave their
runoff/next-highest ranking to ProgA, and two gave it to ModA.
Note that four of the votes going to ProgA were from voters who
originally had ranked ProgC as their first-place ranking, but then ProgC
was eliminated and their vote went to their runoff/second ranking, ProgB,
who also now has been eliminated, so the vote of those four voters goes to
their next runoff choice, i.e. their third ranking, who is ProgA. In
addition, six of the votes going to ProgA were from voters who originally
had ranked ProgD as their first-place ranking, and two were from voters
who originally had ranked MinorA as their first-placed ranking, but with
Prog D and MinorA eliminated the vote of those voters goes to their next
runoff choice, who is ProgA. Here
are the vote totals:
ProgA has passed the winning threshold of 51 votes and is declared the WINNER, with 58 votes. Note that those 58 votes came from: 18 first-place rankings, eight second-place rankings from ProgC voters, 14 second-place rankings from ProgB voters, two second-place rankings from ModB voters, 10 third-place rankings from ProgD voters (six of whom ranked ProgB as their second choice and four ranked ProgC as their second choice), four third-place rankings from ProgC voters, and two third-place rankings from MinorA voters.
Taken together, those votes and the voters who cast them formed the COALITION that overwhelmingly elected ProgA as the representative for this progressive district, with a landslide of 58 votes. Prog A won by having a strong enough core of first-place rankings, but ultimately in such a split field had to pick up many second- and third-rankings from the supporters of other candidates, especially other progressive candidates.
Notice that those voters who ranked either of the top two finishers, ProgA and ModA, as their first-place ranking did not use their second or third place rankings/runoff choices. Thatís because your vote does not go to a runoff choice/lower ranking unless your higher ranking has been eliminated from the runoff. Everyone has one vote in each round of counting, and you donít want to have your vote count for a lower choice if it can count for a higher choice who still has a chance to win. Thatís analogous to the December runoff -- those voters whose candidates made it to the runoff continue to vote for that candidate. Itís only those voters whose candidates have been eliminated from the runoff that pick among the remaining candidates as their runoff choices.
There is no advantage to ìbullet votingî ñ ranking only one candidate ñ since a lower ranking cannot defeat a higher ranking. Itís important to use all three of your rankings in order to participate fully in the runoff, and to allow coalition-building to occur.
You can view a flash animation of the mechanics of how RCV ballot counting will occur by visiting these links: vote and flow.pdf.
The ballot reading and storing of all these rankings is done by the voting equipment in the precincts, the Optech Eagles, which is optical scan equipment (i.e. NOT touchscreens) with a fully voter-verified paper trail (your ballot) and by the central scanner, the Optech IV-C, that scans absentee ballots. After the polls are closed, a software program on a computer will take less than five minutes to count this race. However, if the race is close, final results will depend on absentee and provisional ballots, like they always do in a close race. And that may take up to a week to scan those ballots, at which point the final tally will be run on the computer and final results announced.
Compiled by Steven Hill, (415) 665-5044, [email protected]