Why Single Transferable Vote is Wrong for Internal DSA Elections

Why Single Transferable Vote is Wrong for Internal DSA Elections

There is a debate within the Democratic Socialists of America (DSA) around voting systems that emerged around the time of the 2019 National Convention, and re-emerged with a local focus within Chicago DSA after the results of the chapter’s 2021 Convention Delegate election. In particular, there has been a lot of debate around two methods that are most often used within DSA, the Borda count and Single Transferable Vote (STV), although the use of other voting methods has also been debated. While some proponents of STV may argue that a “big tent” organization like DSA requires a voting system that accounts for a range of minority opinions, I worry that the proportionality achieved by STV can lead to more division and confusion in our internal elections and make it harder for DSA to act as a unified whole.

Different Systems Produce Different Results 

Borda and STV are both forms of ranked choice voting. The difference is how each system “counts” votes. 

Using Borda, every candidate you rank is assigned a certain number of Borda points based on how you ranked them and on the number of candidates in the election. The Borda count for each candidate you rank is calculated based on the number of candidates total minus the position you’ve ranked the candidate on your ballot. So in a four candidate election, the candidate you ranked first would receive three Borda points because you ranked them ahead of three candidates, while the candidate you ranked 2nd would receive two Borda points, and the candidate you ranked third just one point. So if you tell a candidate you will rank them 2nd in a 4 candidate election, you can guarantee that you will be giving them two Borda points. 

Using STV, each voter only has one vote which can transfer partially from your 1st choice candidate to your 2nd choice candidate depending on the different “stages” of an STV count. In STV elections, there’s a “winning threshold” that’s calculated based on the number of votes divided by the number of seats being elected plus 1 — plus 1. So the winning threshold for an election for three seats with 40 voters would be 11. Once a candidate hits 11 votes they would be elected; if they have more than 11 votes, their “surplus” would then be allocated down to other candidates based on how their voters ranked other candidates on their ballots. 

From Opavote:

More generally, counting votes with STV proceeds as follows:

  1. Count the first place votes.
  2. Transfer votes from one candidate to other candidates:
    1. If a candidate has surplus votes (votes in excess of the winning threshold), then transfer surplus votes to their next choices.
    2. Otherwise, eliminate the last place candidate and transfer those votes to their next choices.
  3. If not all seats have been filled, then go to step 2.

Because of this, in STV elections voters can only guarantee that they can give their first choice candidate a single vote; if their first place candidate doesn’t get elected until the final stage, then how they ranked the other candidates is essentially null. Even if your 1st place choice does have a surplus, your second choice candidate will only be receiving a percentage of a single vote. 

The most common voting method DSA members will encounter in elections for public office is First-Past-The-Post (FPTP). FPTP can be used in a multi-winner election, but there is no option for ranking candidates. You would either vote for a candidate or not, and each vote for each candidate would be equal to 1. If there are three seats to fill, under FPTP the three candidates with the most votes total are elected. 

To better understand these three voting methods, let’s look at some real results with each different system and its results used. Linked here is a spreadsheet showing the winners of the 2021 Chicago DSA National Convention Delegate elections using STV, Borda, and FPTP counts. The STV count is the official results while Borda and FPTP counts have been calculated using the raw ballots from the election. Each count produces a different set of winners. 

The most notable difference between the three results is that the STV count elected six candidates who would have not been elected had a Borda or FPTP count been used. For instance, using a Borda count the candidate Dmitri M. would have been ranked 32nd in the overall count and 20th using a FPTP count (having been ranked between 1 and 49 on 52% of all ballots cast)  and elected using either system. Using STV, this same candidate was eliminated in the final stage (although elected as an alternative in the separate count for alternates using STV). Another candidate, Greg G., would have been ranked 63rd in a Borda count or 65th in a FPTP count (having been ranked between 1 and 49 on 20% of ballots cast) and not elected as either a delegate or an alternate. However, using STV the candidate was elected as a delegate in the 54th of 60 total stages in the count.

The Cost of Proportionality 

STV is a voting system that attempts to provide proportional representation in election results. In a highly partisan election, STV allows each party to receive a number of seats proportional to their number of voters. It’s for this reason that STV makes sense for elections for public office. But what about in an organization like DSA? Party is substituted out for internal DSA factions, candidate slates, caucuses, or otherwise organized political tendencies. Again, in a highly partisan environment, allocating seats on a proportional basis makes sense. If Caucus A has 10% of the vote, why shouldn’t they receive 10% of the seats up for election? 

Looking at the ballots from the 2019 Chicago DSA Delegate Elections, only 155 voters (or 36% of the electorate) cast ballots with “partisan discipline” where their top candidates were all from one caucus or slate. The vast majority of ballots were more of a mixed bag with voters’ top candidates consisting of a mix of candidates from across different affiliations as well as independents. While some ballots lean towards one slate versus another, most voters cast ballots in a very non-partisan manner. 

For this reason, both non-partisan candidates and voters are at a disadvantage under STV, as ultimately the cost of proportional representation is electing candidates with strong partisan support (even if they only appeared on a quarter of ballots total) over candidates without strong partisan support despite winning a broad array of support from the overall electorate. 

The 2021 Chicago DSA National Convention Delegate election results give a prime example of how STV can allow a small but disciplined faction to win seats despite not having the support of most voters. The Chicago local of the DSA Class Unity Caucus (CU) ran a total of 12 candidates for delegate, although only one candidate identified themselves in their candidate statement despite the chapter’s internal elections committee encouraging candidates to disclose any caucus affiliations in their candidate statement. Multiple CU candidates wrote candidate statements of one or two sentences, and very few attended the two delegate forums for members. Ultimately, CU was able to elect four candidates as delegates and one candidate as an alternate because of a 36-person disciplined voting block, the vast majority of whom only voted for CU candidates. 

CU’s success in electing four delegates was not a result of winning over non-caucus members to their caucus’ politics. Only CU members would have known exactly which 12 candidates to rank first on their ballot because almost all of their candidates were not transparent about their affiliations. CU candidates didn’t need to write convincing candidate statements or worry about performing well in the candidate forums; they just needed to make sure their caucus members registered and voted, knowing that their caucus numbers alone would guarantee them a proportional percentage of delegates. It’s precisely because of the STV voting method that CU knew they could approach the election in this way, and it is likely why the caucus is pushing both locally and nationally to make STV the standard for internal DSA elections. 

What Holds The Tent Together 

The argument can be made that despite the costs noted above that proportional representation through the use of STV elections is still vital for DSA to remain a “big tent” organization. As comrade Al Gordon wrote in his editorial a few weeks ago:

If CDSA is to be a successful big tent organization, we will have many minority factions, each with little or no likelihood of ever becoming a majority faction. These factions will rationally expect to see their interests represented in our decision-making process, corresponding to the people-power that they contribute to our chapter.

…It would make no political sense for these factions to remain in CDSA and conduct themselves in good faith if they must submit to an electoral framework that is skewed against them. Without proportional representation, over the long run, our minority factions will leave or resort to unofficial channels to seek recourse for their underrepresentation, subverting our chapter’s transparent democracy. The political-economic reality in front of us is that our big tent will collapse without proportional representation holding it up.

The question I have in response to this argument is who decides how big the tent is and who it is open to? Essentially, anyone who pays dues to DSA and broadly agrees with the principles of the organization can be a member, and DSA historically does not have a strong political line or system for disciplining members. The only real limits to membership are found in Article 1, Section 3 of the DSA bylaws

Members can be expelled if they are found to be in substantial disagreement with the principles or policies of the organization or if they consistently engage in undemocratic, disruptive behavior or if they are under the discipline of any self-defined democratic-centralist organization.

Unless DSA shifts to having a tighter political line with stronger mechanisms for discipline, the way members set the boundaries of our “tent” is through internal elections. To be elected using a Borda count or even FPTP members need to have support from a large percentage of voters, allowing members to set the limits of what political tendencies are represented within DSA. The Borda count is particularly useful in allowing electorates to rank candidates from most to least preferred, with least preferred candidates accumulating less points total and thus being lowered in the total count. Using STV, members just need support from a small but disciplined faction, no matter how other voters rank them, and thus members lose their power to set political limits for the organization and instead are forced to bow to any grouping, whatever their politics or their approach to DSA, that can muster the small percentage of voters needed to win representation under STV.  

My main point of disagreement with Comrade Gordon is ultimately his evaluation that “our big tent will collapse without proportional representation holding it up.” In my opinion, what holds our big tent together is our points of unity, shared vision and goals, and our willingness to work together through the organization. Despite some of the internal conflict that occurs within DSA, the organization’s major caucuses and organized tendencies share commitments to the labor movement, to thoughtful engagement in elections, and in winning transformative policies like Medicare for All, and understand that all of these commitments require mass organization. I believe that the much greater threat is if the proportional representation produced by STV leads to a DSA made up of tendencies and factions that are unwilling to work together because of political differences perceived to be irreconcilable. 

I don’t believe DSA should be an open house where any self-described leftist is guaranteed a seat at the table; I believe DSA must be a united front of the working class building power together to achieve a common vision. The latter organizational form leaves room for debate on strategy and diversity of political ideologies within the anti-capitalist tradition, but it is also a form that depends on the coalition that makes up DSA being willing to work together and work towards achieving shared goals. If our organization continues to use STV or enshrines it into our national bylaws (as is proposed at this year’s national convention), I worry that we will run the risk of moving away from being a big tent mass organization and towards becoming a shared common space where different sects compete for representation in an organization where members are focused primarily on contesting power internally on behalf of their faction rather than externally for the whole of the working class.