The Board of Elections is taking way too long to tabulate ranked-choice voting results
Election officials in New York City are unnecessarily withholding crucial election results from the public regarding the races for mayor, comptroller, City Council and other offices. Specifically, they have refused to run the ranked-choice voting (RCV) tally for the 800,000 ballots that, as of this past Friday, were in their possession.
That has fed a false perception that RCV is so complicated that it is causing delays in election results, when in fact the delay is being caused by a misguided Board of Elections policy. In this age of controversy over vote suppression and claims of stolen elections, having hundreds of thousands of uncounted ballots lying around for a week can only feed suspicion and cynicism.
Especially since it’s completely unnecessary. I am the main architect of the ranked-choice voting system used in San Francisco, Oakland and other cities. In the first RCV elections in those cities, election officials made a similarly flawed decision. Like in New York, administrators erroneously assumed that the RCV tabulation would be operationally complicated, and that public confusion would ensue if the outcome changed between an initial RCV tally on election night and the final tally many days later. Especially with so many absentee ballots to count post-election, administrators thought it better to wait until most ballots had been processed.
But the dynamics turned out to work just the opposite. In the 2010 mayoral election in Oakland, election officials waited a number of days before running the first RCV tally. The media and the public began assuming that the frontrunner (based on first-choice results) would win. When the second-place finisher came from behind to claim victory, chaos ensued. That undermined the credibility of the election results and the legitimacy of the new mayor, who happened to be the first Asian-American woman elected mayor of a major U.S. city.
If the election officials had run the RCV tally on election night, the public would have seen right away that the frontrunner was actually losing an extremely close race. By waiting, they caused the very confusion they had feared.
Having learned their lesson, now the election officials in San Francisco and Oakland run the first RCV tally on election night, and then another one on each day following the election. It’s easy and fast to do. They tally a data set composed of all ballots processed up to that time, and announce the results as “preliminary” (as even non-RCV elections must do). They also announce how many absentee ballots remain outstanding to count.
In addition, the Bay Area officials post on the internet a digital text file of the “ballot images,” which is an anonymous record of each voter’s rankings. That allows the media and members of the public to download the ballot images and, on their own computers, run the RCV tally. This has created maximum transparency, and established a precedent for crowdsourced election security.
Doing all this actually fosters clarity and eases the concerns of the public. If only first choices are initially released, as in New York, nobody really knows what is going on in that election, especially if the race is close.
Operationally in New York, the memory sticks holding digital ballot data from the 4,076 ballot scanners in polling stations are required by contract to be transported to election authorities within three days of the election. So as of this past Friday, all memory sticks were in their possession. It takes a few hours to upload the data from the memory sticks into the central counting tabulator. Then, in a matter of minutes, the computer does all the counting and transfers from losing candidates to remaining candidates.
Note that, with a Friday deadline, a great many of the 4,076 memory sticks likely had been delivered on Wednesday and Thursday. Even with the outdated New York State law that prolongs the processing and counting of absentee ballots, by late Thursday afternoon election officials likely had enough ballots to run an initial RCV tally. With that information, the public would likely have known if Eric Adams’s lead had increased, or whether Maya Wiley or Kathryn Garcia had caught up or overtaken him. The public would have received a reliable snapshot regarding the closeness of each race.
Instead, everyone will be in the dark until next Tuesday, a whole week after the election.
The idea that RCV has slowed down the election results is false. It actually takes very little time to run the re-sorting algorithm on a computer. While I understand the caution on the part of the Board of Elections officials for their maiden RCV election, what they must realize is that their excessive caution actually creates even more problems. Sequestering ballots and withholding RCV tallies can undermine public confidence.
The best practice is to run the RCV tally early and often, and to release the ballot images to the public for crowd-sourced verification.
Steven Hill is the author of seven books, including “10 Steps to Repair American Democracy and Fixing Elections: The Failure of America’s Winner Take All Politics.”