We were encouraged when we heard that FairVote was updating its online article on STAR Voting in light of concerns from the Equal Vote Coalition and a number of leaders in the electoral reform community. However, in reading the updated article (July 2018) we found it disappointing. They dropped some problematic assertions while retaining others, but overall it fails to address the central concerns raised against it in the first place.
The article is likely to deter non-specialists from supporting STAR Voting, even though the arguments do not hold up to critical scrutiny. Our response to the original article is still available and goes into great detail, but here we will limit ourselves to summarizing three fundamental problems with their article.
Problems in FairVote’s Critique
1. Puzzling recommendation. FairVote writes:
“It is difficult to know for sure whether STAR voting can break this tendency of [tactical voting in] scoring voting methods. It’s lack of track record in meaningfully contested elections where voters care about the results, candidates run serious campaigns, and the results are publicly revealed afterwards in a way that might affect future tactics [leaves us with insufficient data].”
The logical conclusion from the above quotation is that STAR Voting should be tested in real public elections in order to confirm or disprove FairVote’s tactical voting concerns. Local municipal and county elections are the venues where such conditions would apply and provide the data to determine what tactical voting concerns, if any, play out in the real world. In spite of this, FairVote concludes that STAR Voting should build a track record of use in private associations where none of their criteria will be met nor the pertinent data gathered.
2. Failure to address rebuttals of their “tactical voting” argument. FairVote’s principal argument against STAR Voting is a hypothetical example of how voters might use tactical voting. We have repeatedly debunked this argument, not only in our response to their article but in earlier email exchanges, yet both their original and revised articles repeat this flawed line of reasoning.
Their argument is that voters might vote tactically to keep a strong challenger to their favorite candidate out of the automatic runoff by giving candidates they like less a higher score than the challenger. FairVote relies on two examples to illustrate this point, neither of which involve STAR Voting.
STAR Voting’s automatic runoff feature makes this tactic very risky – more likely to give such voters a worse outcome than if they had voted according to their true preferences. The idea that voters would employ a complex and risky strategy that is not likely to benefit them on a wide scale is contrary to both common sense and FairVote's own data. (This data has shown that voters’ desire to vote honestly often overcomes tactical considerations even when an honest vote actually could give them a worse outcome, as can happen with Ranked Choice Voting.)
3. Biased use of criteria for evaluation. Election scientists evaluate voting methods against a wide variety of criteria, some of which conflict with each other. Yet FairVote bases its assessment of STAR Voting primarily on a single criterion – Later No Harm – which is about the method’s incentives or disincentives to give honest support to candidates beyond a voter’s favorite.
We find it particularly conspicuous that FairVote doesn't mention that Later No Harm is only one half of a paradoxical set of important criteria. The second criterion in that pair is Favorite Betrayal, which considers whether or not it is safe to fully support your favorite candidate in the first place. FairVote’s preferred voting method, Ranked Choice Voting, fails Favorite Betrayal in predictable ways. STAR Voting deliberately balances the two critical criteria of Later No Harm and Favorite Betrayal to minimize the effects of each.
"Unfortunately, satisfying both criteria is impossible without introducing other undesirable qualities to the voting system… Whether they’ve realized it or not, folks who tout Later No Harm as the holy grail of voting systems criteria are actually saying that The Spoiler Effect is not a problem they think is important to fix." (Dempsey, “A Farewell to Pass/Fail: Why We Ditched Later No Harm.”)
Choosing to satisfy Later-No-Harm over Favorite Betrayal opens the door to the same pitfalls we are trying to solve in our current voting system: “Lesser-Evil” voting and “Spoiler” problems in a system plagued by vote splitting.
4. Any many more… This failure to address counter-arguments persists throughout their article. You can find our full point-by-point analysis of their original article here.
Conclusion
We respect the role FairVote has played in leading the charge on voting method reform in the United States. In many ways they have laid the foundation for the current renaissance in election reform. However, FairVote's recent article stands in contrast to their stated claim to be neutral on STAR Voting and supportive of better voting methods in general. Rather, FairVote's track record shows a pattern of working over decades against voting method reforms other than Ranked Choice Voting.
We applaud them for making a huge step in the right direction by taking down a number of these "attack" articles recently, but urge further action. STAR Voting was created as a hybrid of features from both Ranked Choice and Score Voting to address criticisms of each while maintaining their benefits. As a win/win method, we hope that STAR Voting can help unite the disparate camps of the voting reform movement.
As this is one the most critical issues of our times, we need to work together to succeed.