Statement on East Bay Delegate Elections
The upcoming DSA National Convention comes at a moment of historic opportunities for our organization and the multi-racial working class movement as a whole.
While there is a great deal to be excited about, and we fully expect the convention to be a huge success, we were disappointed to hear that several members of DSA are challenging East Bay’s election of delegates to the DSA National Convention. As the co-chairs of East Bay DSA, we want to address that challenge directly.
The challenge rests primarily on the claim that Borda count — the system of ranked-choice voting we used for our elections — is an illegitimate method for electing delegates to the national convention, and that the Single Transferable Vote system would be preferable. In this response we address the question of why we chose Borda count first and then address several of the other concerns raised.
We hope that DSA members will uphold our election and reject the credentials challenge — though as we note below we agree with the challengers that DSA should have a standardized election system for national delegates across the country. We would support working with the challengers to write an amendment to the DSA constitution along those lines. And as we outline at the end of this statement, there are certainly important changes we can make to how East Bay runs elections to make them more efficient and more clear. We plan on carrying these changes out this year, and hope other chapters will also do the same.
Why did we choose Borda count?
For at least the last three years that we have been members, East Bay DSA election methods have been decided by chairs. In 2017, and in prior years as far as we know, East Bay used a simple first past the post system for electing national delegates. After our chapter convention in May, our outgoing co-chair Zach M. and reelected co-chair Abi had to quickly pick a voting system for the national delegate elections. Like most DSA members, none of us claim to be experts on the use of different systems for electing delegates, so we initially planned on using the same system as was used in 2017.
Three weeks before the start of the delegate election, some of the writers of the credentials challenge contacted Zach and Abi to request that we switch to ranked-choice voting. We were convinced by their arguments that a simple first past the post system would not ensure that all tendencies would be represented at the national convention. We respect and support DSA’s big tent character and believe all points of view should be represented.
Zach was not running to be a delegate so he took the lead to research options for ranked-choice voting. As anyone who has looked into the options knows, Zach found that every ranked-choice system for multi-winner elections has serious flaws in the context of a 63-candidate, 38-winner election.
At the time we were particularly concerned that STV seemed to be biased towards the most well known figures in a tendency because the person who receives a certain number of first place votes is almost guaranteed to win, despite the fact that they might not have a broader base of support. We were concerned that explaining and understanding the results from STV elections can be very complicated. And we were concerned that STV breaks ties randomly — meaning that you can run an STV election multiple times with the same ballots and get significantly different results, essentially meaning candidates can win or lose an election due to a coin flip.
As an alternative, Zach proposed the Modified Borda system. Both Zach and Abi felt this was the best choice given the options and the fact that we were running up against the deadline to elect delegates and had to act fast. Our understanding is that Borda is also a system of proportional representation, but one that gives candidates who are less high profile but have broader bases of support in a chapter a better chance of winning. We also learned that other chapters had used Borda count for national delegate elections and that Borda count had been used by some branches in New York City for their city convention election. Most importantly for helping us make our decision, we learned that national DSA had used Borda count for the election of the 100 at-large delegates (at-large delegates are delegates who do not live in the area of an officially-recognized DSA chapter).
Were there other flaws in the election?
We absolutely reject the insinuation in the challengers’ letter that there were other improprieties in the election.
The challengers argue that the modified Borda system we used was particularly flawed. We do not believe this is the case. There is a version of Borda that allows voters to game the system by “bullet voting” — that is, just voting for a few candidates so they get the most points. We wanted to avoid this so we used the Modified Borda system that disincentivizes bullet voting.
We also ensured that the votes were counted by members from many different tendencies, none of whom were running to be delegates.
Unfortunately it took us several days to count the votes. In the past our chapter has not used online voting, so we did not use online voting for this election. In hindsight we recognize that that was a foolish decision, as it took several days to count so many votes. In the future we will support switching to online voting for big elections.
Why not rerun the election using STV?
Some have suggested that we should rerun the election using STV — using the same ballots cast last month. The flaw with this is that voters were given instructions on how to vote using Borda count, not STV. It would therefore be impossible to fairly rerun the elections using STV, since had we used STV voters would have ranked candidates differently.
Why were the results different between our chapter convention and the national delegate election?
The challengers note that a higher proportion of members of one tendency were elected in the national delegate election in June than were elected to our chapter Steering Committee in May.
This is not a legitimate challenge. These were two different elections, for two different kinds of offices, with two different sets of candidates, and two different electorates. For example, in the May election there were 280 voters, but in the national delegate election in June there were 345 voters. Moreover, while the Steering Committee election was very clearly about different slates running against each other (and due to the significantly smaller number of candidates it was much easier for voters to figure out who was on which slate), the national delegate elections were not nearly as clear. Many candidates did not note the tendency or slate they were running with in their candidate bios. Many voters likely voted for the candidates they knew best from organizing, regardless of what tendency they are members of.
Why not release the individual ballots?
Raw ballot data has been released in two East Bay DSA Steering Committee elections but was not released in the previous national delegate election.
We decided not to release individual voter’s ballots because we had serious concerns that doing so would compromise our ability to guarantee that all voters would be casting secret ballots. Since a significant number of candidates in our national delegate election received only one first place vote (almost certainly on their own ballot), many voters’ ballot choices would be easily identifiable. This concern was raised by multiple delegate candidates to Steering Committee members before and after the election. Members who asked about this were assured beforehand that the individual ballots would not be released. The same concern does not apply to Steering Committee elections because there are far fewer candidates, and candidates receive more than one first place vote.
However, we as co-chairs of the chapter agree that we could release vote count data that does not reveal individual ballots and how each ballot was organized. We will talk to our fellow members on the East Bay Steering Committee at our meeting on Sunday and propose next steps.
What comes next?
The challenges we faced in picking a voting system were hardly unique to our chapter. In talking to members of Steering Committees across the country since we learned about the credentials challenge, we’ve heard from many members that choosing between the different systems and their variations has been difficult.
Moreover, many chapters — not just East Bay — did not use STV. So far, we know that Fresno and Louisville DSA used Borda count, national DSA used it for at-large delegates, and we suspect other chapters did as well. Seattle and Los Angeles (and likely many other chapters) did not even use a system of proportional representation — instead they used a simple first past the post election in which candidates voted for as many candidates as there were spots for delegates without any ranking system.
We believe it would be egregiously unfair to unseat members of the East Bay delegation because they were elected using a system other than STV while many other elected delegates who were similarly elected by systems other than STV keep their credentials. To be clear, we do not believe any duly elected delegate should have their credentials revoked. The DSA constitution leaves it to chapters to decide the method of election of national delegates. Until that is changed, we have no choice but to respect the election methods chosen by chapters.
We would however support an amendment to the DSA constitution to change this in the future. We — and we know we are not the only leaders in a DSA chapter who feel this way — would welcome a standardized election process for national delegates for all chapters. We would prefer not to have to try to understand and explain the intricacies of Borda count, STV, Hare count, and all the other varieties of electing delegates in the future. A standardized recommendation from the national office, informed by people with some degree of expertise in picking voting systems, would be ideal. It seems like a consensus is emerging in DSA that STV is the preferred system. We think it would be desirable for national DSA to consult with experts on this question and issue a general recommendation to chapters about what the best practice should be. We would be more than happy to adopt STV for future elections if the conclusion is that that is best.
However, we urge delegates to the 2019 National Convention to reject the current challenge being made to our delegates’ credentials. And we urge the challengers to be principled in their approach and not hold East Bay to a double standard. If they really believe STV is the only legitimate system for electing delegates, and all other systems are so deeply flawed as to invalidate the elections, they should challenge the credentials of all delegates elected by other systems, including all 100 at-large delegates who were also elected by Borda count.
We are looking forward to the convention and hope it will be an inspiring and unifying experience for all members. We hope that the beginning will not get off to a rough start with a confusing credentials challenge. Duly elected delegates from East Bay, like every other delegate, have bought tickets, taken time off of work, arranged childcare, and made other sacrifices to attend the convention. Stripping them of their credentials at this point would be manifestly unfair and arbitrary. Many of us attended the 2017 National Convention and were surprised and frustrated by attempts by a small number of members to remove the delegate credentials from the NYC delegation. We stood with our NYC comrades and a huge majority of national delegates to uphold that their elections were fair and they deserved to attend the convention. We hope comrades will similarly support our delegation.
Finally, going forward in East Bay we commit to helping ensure that changes are made to address the legitimate and comradely concerns brought by the authors of the challenge.
First, the voting method must be announced weeks in advance. We apologize that this wasn’t done this time — though we ask that members understand that as volunteers with busy day jobs we are not always able to act as quickly as we would like.
Second, we believe the whole Steering Committee, not just the co-chairs, should decide voting methods going forward.
And third, we commit to supporting a switch to online elections using Opavote in the future. It was a major waste of volunteer time to tally ballots for several days.
In solidarity,
Abigail G-G and Keith B B
Co-chairs, East Bay DSA