This is our updated annual post that makes the case for why giving to The Center for Election Science is an awesome decision. I’ll give a background on our organization, include our updates for this year, and provide some key information to help you with your giving decision. Links in the article will tend to go towards more technical articles on our site. Feel free to ask any questions you have.
Our organization studies voting methods and works with chapters across the country to implement approval voting. This is important because the voting method we use now is terrible and approval voting does a much better job at winner selection, measuring candidate support, and other factors that make a voting method good. And it does this better than alternative options. The way trillions of tax dollars and government policies are decided are at stake.
We’ve quickly implemented approval voting in two cities already and we now have a campaign in Seattle that has grown out of our chapter system. We’re also doing a large-scale polling project to focus on states next, which will let us target federal offices, their primaries, and presidential electoral votes.
Given that much of existential risk is made up of human-caused events stemming from issues that can be mitigated by policy and better governance, this intervention is even more important. This intervention via ballot initiatives is both efficient and costly at the scale we’re working. Over a 6-year period, we have a budget gap of $250M, though if that funding is accelerated, we may be able to shorten part of our timeline. Using ballot initiatives at this scale may also set a model for other interventions. Because this is a technical area, the FAQ covers many common questions.
- What we do
- Why we do what we do
- What we’ve done and what we intend to do
- Why we need funding
- Why there is urgency
- Benefiting the future
- Our ask
What we do
We empower people through voting methods that strengthen democracy. We do this through research and advocacy with a focus on approval voting.
On the research side, we collect data comparing different voting methods. We publish articles on our blog as well as in peer-reviewed journals, which we make sure are open access. For example, we published an article comparing the impact of different voting methods on the 2016 election. Earlier this year, we hired our Director of Applied Data & Research to enhance our research capability. We’ll be sharing more research in the future.
On the advocacy side, we work with local groups to run educational campaigns alongside ballot initiatives. We give local and state-wide groups the tools they need to involve their communities so that they can implement better voting methods. We do this through a national chapter system that we developed this year. We focus heavily on approval voting due to multiple factors but mainly its simplicity and strong performance.
Why we do what we do
Virtually across the globe, we all use the worst voting method there is—a choose-one voting method—to elect people to executive and other offices. We trust those same people we elect using that terrible voting method to (1) spend vast sums of taxpayer money and (2) execute policies that impact our lives.
Our current choose-one voting method is bad for our democracy. Choose-one voting causes vote splitting between candidates, often creating an inaccurate picture of voter preference. This vote splitting causes spoilers and can squeeze out moderate candidates. Good candidates sometimes don’t even run for fear of being labeled a spoiler, or because of manufactured notions of unelectability. This choose-one approach also incentivizes voters to vote against their favorite candidates if they perceive them as not being viable. In all, this results in inferior government and a poor environment for fresh ideas to be discovered and implemented. It also means a lack of continuity with policy due to wild ideological swings from administration to administration.
There are much better ways of electing these people to office, one of which is very easy—approval voting.
Approval voting is classically a single-winner method that lets voters choose as many candidates as they want. The candidate with the most votes wins. This voting method tends to elect more consensus-style candidates while yielding an accurate reflection of support for all candidates on the ballot, not just the winner. With approval voting, voters can always vote for their favorite no matter what.
These features produce elected officials that have the support of their constituents and a more stable political environment less susceptible to wild ideological swings. This elevated level of support can permit new ideas to develop that otherwise couldn’t under our current choose-one method. Consider that presidential candidates aren’t even allowed on the debate stage to voice their message without polling at least 15%—a feat nearly impossible under our current method for anyone outside the two parties. The approval voting winner also tends toward the middle voter and electorate consensus—which should not be confused with an unopinionated candidate.
What we’ve done and what we intend to do
Over the past three years, we have helped make approval voting a reality. We brought approval voting to its first US city in 2018, passing an initiative in Fargo, North Dakota. Our victory was built on collaboration. In partnership with a local organization, Reform Fargo, our public education campaign helped pass an initiative that was seen as a long shot by local political insiders. The ballot initiative passed with 64% of the vote. All of this was achieved within one year of our initial funding from Open Philanthropy.
After the passage of approval voting, Fargo, a city plagued by vote splitting, elected commissioners in 2020 with over 50% approval with approval voting. This clear support was a first in recent memory.
In 2020, we brought approval voting to its second city, St. Louis, MO. Again, we worked with the local community to make this campaign a success. We partnered with organizations including the League of Women Voters and the local SEIU affiliate, among others. The ballot initiative—Proposition D—passed with 68% of the vote. In just two years we replicated the success of Fargo’s implementation and scaled it by over twofold with St. Louis’ population of just over 300,000 people.
In 2021, St. Louis used approval voting for the first time. Unlike previous elections where St. Louis experienced severe vote splitting, particularly among Black and progressive candidates, this election produced winners with consensus support. The old process elected a mayor who doxed protesters in the city. Immediately following the passage of approval voting, this incumbent mayor decided not to run for reelection.
With approval voting (using a top-two system with approval voting in the primaries), the next election went differently. Two candidates advanced to a runoff, with the leader having 57% of the vote. From a geographical perspective, the early leader and eventual winner, Tishaura Jones, received support on both sides of the traditional racial divide in the city, demonstrating the consensus-building effect of approval voting.
In November of 2021, Seattle Approves launched its campaign to bring approval voting to a new jurisdiction. Seattle has a population of over 750,000 and is the 18th largest city in the United States. Polling for this initiative is already at 70% out of the gate.
Our Director of Campaigns has also set up a chapter program that has organized groups across the country. Those chapters, in combination with strategic resources and a request for proposals for funding, will allow us to take a tactical approach to campaign planning in the future.
Our Director of Applied Data and Research has also led a first-of-its-kind polling project that will amass an unprecedented dataset to guide our state-by-state path. In the near future, our goal is to pass initiatives at the statewide level across the country. This escalation means simultaneously implementing approval voting at the federal level and in primaries, including US Senate, US House, and US presidential electoral votes.
Why we need funding
We have been extremely grateful for the funding we received from Open Philanthropy. And we are grateful to Will MacAskill for believing in our ability and prospective impact enough to recommend us to Open Philanthropy. Our past support from the Long Term Future Fund as well as Survival and Flourishing have also added to that foundation for our early success.
To succeed in the next phase of our mission, we need to scale up our funding as we begin to target statewide efforts. Our future plans require having the staffing necessary to coordinate a nationwide effort of multiple concurrent state-wide campaigns while simultaneously doing the necessary research to ensure a data-backed approach.
We estimate this will require an annual budget of $15M. Our current annual budget is only $1.5M. Alongside this funding, we will need to provide financial support for the campaigns through a sister 501(c)4 organization, which does not have a cap on campaign spending. We estimate that we will require at least $250 million in funding over the next 5-6 years to cover all the ballot initiative states and maintain the infrastructure to run this scale of operations.
This number may sound high—and it is a large number—but it’s efficient when you consider the trillions of tax dollars and the scope of the policies at stake. Also, when you consider that the cost to gather signatures necessary for an initiative to merely get on the ballot in a large state like California is $10M, the cost makes sense.
It’s important to note that our work may offer benefits beyond approval voting—though approval voting is the core and most critical impact. In talking to other EA’s, they’ve pointed out that this methodical approach can serve as a blueprint for other EA organizations considering policy change through ballot initiatives. The entire EA community may benefit from our challenges and successes.
Overall, we have an enormous budget gap. We’re creating a large opportunity to fund a movement with systemic impact. We’ve already demonstrated proof of concept and replication while also scaling. Now, we’re ready for the next phase.
Why there is urgency
Achieving voting method reform through legislation is a nonstarter due to the conflict of interest for legislators elected using existing methods. Instead, ballot initiatives are the main tool for pushing approval voting. Only about half the US states permit ballot initiatives. Even fewer have the legal framework that permits initiatives to change voting methods at the city or county level.
Unfortunately, we have already missed opportunities because of our funding limitation. Prospective partners have passed us by for the sole reason of lack of funding, even though they prefer approval voting to other reform options. We have to act quickly to reverse this trend. There are other groups that are open to approval voting and even prefer to work with us. But they will go with the heavily inferior RCV if they perceive that we cannot provide appropriate resources. This means our funding and rapid pace have to continue. This does create some tension .
Given how close approval voting came to matching RCV’s momentum in 2020, it’s clear that with funding, approval voting can move faster than RCV. Tens of millions went into RCV campaigns in 2020, supported by organizations with budgets averaging over $4M. In contrast, we had a $1M budget at CES. If that funding discrepancy increases, opportunities for approval voting will get displaced. And we will have offices voted on using an inferior system that doesn’t adequately address the fundamental problems that need to be solved. Switching to approval voting in the future may also be harder to do in places where RCV is used now.
Aside from displacement, we’re seeing the critical nature of electing the right people to office—the pandemic providing the most salient recent example. We are at a crucial point for the life of our democracy, one that requires bold commitment to save it. Approval voting needs to be implemented as quickly as possible.
Benefiting the Future
Approval voting is about the present and the future. This is based on the simple premise that a stable political system populated by elected officials with the consensus support of their constituents will improve policies and government over time. Surely, our current system is not getting us there, and neither are alternatives being proposed.
We are building the case that focusing on approval voting and thus policy will affect the far future. This is because approval voting is more stable compared to other methods and it allows new ideas to grow, some of which are crucial for our long-term prosperity. We have successfully made this argument to support past grants from the Long Term Future Fund as well as Survival and Flourishing.
This is not to say that approval voting accomplishes these long-term goals completely independently. Rather, it creates an environment where stable and responsible policy can take hold. And doing so is time-sensitive because it can take time for policy and systems to change.
According to Toby Ord in his book, The Precipice, the probability of an existential event over the next 100 years is one in six. That number itself is alarming and should get our attention. But it’s worse if you extrapolate the risk over time. For example, if that risk remains a constant, it implies that the probability of an existential event a mere four centuries down the road will be roughly a coin toss.
It’s also worth noting that the substantial existential risks that he identifies are human-caused and affected by policy and government. Looking at the numbers, it becomes critical that we identify multiple approaches to bring this number down. Using approval voting to improve our government is one such approach.
Many of you are already familiar with our work. Maybe you heard about the Open Philanthropy grant or its renewal in 2019. You may have heard us speak at EA Global or REACH Berkeley. Or perhaps you listened to the 80,000 Hours episode. Regardless, our team and I are grateful that you take this issue seriously enough to direct your attention to it.
And to those of you in the EA community who have already donated, thank you. It takes a special kind of donor to support this bold, analytical cause. Many of you have reached out to us following your gift to let us know about your interest in this effort.
Please consider a gift that matches both your capacity and your commitment to fundamentally improving government. You can give online through our website. To give by other means or at a significant level, please reach out to me personally. I will be happy to get back to you, even over the holidays. We are a sophisticated team so we can also handle complex assets if you let us know your situation. You may have already seen many of my essays on philanthropy, advanced giving instruments, and tax consequences on my personal website.
Thank you again to those in the community who have already given.
 It’s important to note that we do not oppose RCV campaigns that are off the ground. We do critique different voting methods and form opinions about them because we value making sure that there is a clear assessment of the options. But we do not tell people to vote against an RCV measure once a campaign has started, and nor do we do opposition messaging to target active campaigns.
The same has not always been the case the other way. The executive director from an RCV group in Washington is publicly opposing our Seattle campaign, and an executive director for a national RCV group opposed our St. Louis campaign. There, the executive director falsely claimed that approval voting would disadvantage women and people of color in St. Louis. (Recall that after approval voting was implemented in St. Louis that it advanced two women and the woman of color won.)
Q: I heard there was this thing about approval voting that wasn’t so good or that another voting method was better. Also, don’t forget about Arrow’s Theorem.
A: All voting methods have quirks, but we maintain that the quirks of approval voting are comparatively mild compared to the alternatives. You can see this article where we go into all the details about approval voting critiques. Also, I talked with Kenneth Arrow personally for an hour, and he said that our choose-one voting method was bad. Really.
Q: How does IRV/RCV match up to approval voting?
A: Not very well. From encountering avoidable anomalies to being needlessly complex, IRV falls well short of what approval voting can offer. Here’s an article on that topic. And here’s an article on the limits of RCV (which includes its limited merits). Note that IRV and RCV are the same, though RCV is the name commonly used by advocates. This is notably confusing not just because of the multiple names but because there are many different ranking methods.
We have shown repeatedly that approval voting better captures public support compared to RCV. We did this by polling Democratic voters in the primaries using multiple voting methods and compared the results to control measures, both earlier and later in the primaries.
Q: How do you decide what makes a voting method good?
A: We look at the type of winner it tends to elect as well as practical issues from simplicity to implementation. Here’s an article on that topic. Here’s an in-depth article comparing methods using these broad criteria.
Q: Will approval voting increase the number of parties?
A: Probably, but likely not by much. Those parties can, however, get their voice heard (and ignored if they have bad ideas). Here’s an article on Duverger’s Law. (Fun video here). Also, third parties and independents clearly benefit from approval voting. Note that the multi-winner proportional version of approval voting would encourage more parties. But the math is more complicated. We’re focusing on single-winner methods first because we think it will be easier.
Q: Why don’t you go after organizations that do achievement awards?
A: We do, though we limit our resources to high-impact opportunities. Here’s an article about how we worked with The Webby Awards. We’ve also done an article on giving games. I’ve personally encountered some resistance when talking with some large awards organizations. They don’t collect the data to know whether their current voting method is bad. Plus they likely perceive that changing their voting method may reveal that their previously given awards have less value.
Q: The Electoral College is awful. Why aren’t you working to get rid of it?
A: The current actions using an interstate compact to make the electoral college moot would still leave us with that awful choose-one voting method. Approval voting would work with this current approach though (RCV wouldn’t). We wrote a whole article about it.
Q: Why don’t you go after primaries? You should be going after primaries.
A: We go after primaries. We do this at both the local level and in the future will do this at the state and federal level. We target both the general and primary elections. We did that in St. Louis. Local elections in Fargo do not have primaries. We’ve written lots about primaries. Here’s an article. Here’s one, too. Here’s one more. We’ll likely write another one before too long as well.
Q: Why don’t you target third parties to get their support?
A: We target third parties to get their support. Green and Libertarian chapters in multiple states support and use approval voting. The Libertarian Party even uses it for national internal positions. Other third parties use it, too. Many of those folks have already bought the notion that RCV will help them, so we have to explain how approval voting would be better. RCV gives little additional support to third parties. Just look at the minuscule support third-party candidates got in Maine. We’ve also demonstrated experimentally that approval voting is far better for third parties than RCV.
Q: I listened to the 80,000 Hours Episode, but I felt that you didn’t go into enough detail in certain areas.
A: It seems like you always think of things after the fact. Here are some quick follow-up details into areas like voter turnout where I could have given a more complete answer.
Q: How can I help again?