I’ve often been skeptical that US political engagement was an effective use of time for EAs. During and after the 2016 election, I heard people front the idea that defeating Trump might be an effective use of EA resources. I’m skeptical that this is true, and I think it’s easy to fall into the trap of “this thing my social group thinks is good: maybe it’s also the most effective thing by EA standards”. Politics is driven by tribalism, so I think this is especially a risk here.

Recently, I’ve surprised myself by coming to believe that donating to candidates who support policies which reduce existential risks is probably the best passive donation opportunity for US citizens. The main reason I’ve changed my mind is that I think highly aligned political candidates have a lot of leverage to affect policies that could impact the long-term future and are uniquely benefited from individual donations.

While I don’t think that the work of individual US congress members is more effective than the work of organizations like the Alignment Research Center working directly on long-term problems, I think that the presence of large funders willing and able to fully fund organizations working on long-term causes makes supporting political candidates with aligned values a more promising target for individuals donations, since congressional election campaigns are limited in how much funding they can accept from any individual donors.

I think there are more effective donation opportunities but they require special knowledge that the major EA orgs don’t have access to. For example, I’ve been looking for promising aligned people or projects in the infosec space that could use funding to jumpstart their career or project. Since I have special knowledge / expertise here, I expect these are among the highest impact donations I can make.

However, I often get pretty busy and don’t have time to look for neglected funding opportunities. Given time and attention constraints, I think donating to political candidates with a strong commitment to long-term oriented policies is my best default. This year, nearly all my EA donations are going to political campaigns. I wouldn’t have predicted this last year!

Why do I think this is effective compared to other donations?

  • Large longtermist donors - Open Philanthropy, FTX, Jaan Tallinn, etc. - can and do fund most promising organizations working on long-term risks
  • Political campaigns are limited by the size of individual donations from US citizens because of campaign finance laws.
    • Contributions to Congressional candidates are limited to $2,900 per election, so $5,800 per year. (A primary election counts as a different election)
  • I think having candidates in Congress willing to sponsor legislation on long-term issues like biosecurity and AI existential risk could significantly improve the prospects for policy interventions in these spaces
    • There are officials in US government who prioritize long-term concerns, but no elected officials meet this bar
    • The 0-1 difference in Congress is large! This is because a single congress person can sponsor legislation. As a start, it would be very good to have two candidates, one from each party, in both the senate and the house. A common concern among EAs is that supporting candidates might polarize important cause areas. Supporting candidates from different parties could help mitigate this risk and work towards bipartisan support of global risk reduction, an area that should appeal to people of any party.
    • Example policy area I think is high impact: Banning gain of function research. This is a policy that nearly everyone working on long-term issues would like to see happen, but there has not been sufficient political will. I think having a few Congresspeople prioritize this issue & propose legislation could lead to a real ban, or at least a ban on federal funding for GoF.
    • In order for a legislator to be effective for long-term issues, the following must be true:
      • They must be be able to create / identify legislation that would be net positive for the long-term (or have staffers that can reliably do this)
      • They must be able to increase the chance that legislation is passed
    • I think the first premise is true in a number of cases, especially around shifting funding away from dangerous biology research. It’s much harder in the area of AI policy, but I expect there are some pretty net positive policies that could be passed.
    • I’m less confident about the second area, but I think the 0-1 effect is probably large. I’d love some concrete examples or numbers on this.

Uncertainties

  • How much will additional funding help candidates win?
    • Seems highest impact to support candidates where the race is close. If they’re likely to win or lose regardless of donations, that seems obviously not very effective
    • I don’t know much how campaign spending relates to election outcomes
  • Who is vetting candidates for their commitment & effectiveness championing policies which will make a difference on long term issues. Will they do a good job?
    • Personally, I’ve been impressed with Guarding Against Pandemics, who have been vetting and publicly endorsing candidates who might support long-term oriented policies
  • Is it ethical / a good norm to support candidates in other states?
    • I think it’s ethical given current laws and norms (see below), but still have some uncertainty here

Ethics of inter-district donations and the integrity of political candidates

One question I’ve considered is whether it’s ethical to donate to political campaigns in districts I don’t live in. The US allows candidates to receive support from outside their district, and this information is available to the public. Furthermore, it’s not unusual for congressional candidates to get a large share of their campaign funding from outside their district. Given it’s both transparent and a common practice, I think it’s ethical to support candidates in other districts. When thinking through this, I asked myself “would I think it’s okay for those who support (or oppose) abortion rights to donate to political candidates who prioritized the issue, even outside of the donor's district?” and noticed that that scenario seemed fine to me.

It’s important to me that political candidates I support have integrity. I wouldn’t want to support a candidate that didn’t plan to do right by their constituents. I don’t think there’s any significant conflict between supporting policies to mitigate existential risk and looking out for shorter-term concerns. I think there will be some tradeoffs in terms of how they would allocate their attention, but I expect the tradeoffs to be pretty small given how huge the federal government is. I also think upping our chance of surviving the next century is pretty aligned with the interests of voters, and I’d love to see candidates who appealed to voters on this axis.

I don’t think long-term issues like AI or biotech existential risks are fundamentally partisan issues. I’d be happy to donate to high-integrity candidates in either political party who are committed to mitigating these risks. Currently, some issues like Gain of Function research are somewhat split along policy lines, and it doesn’t seem to be for principled reasons. I remember researching SARS-CoV-2 origins in early 2020, and at the time Trump’s support for the lab origin hypothesis caused a lot of people to dismiss it as a conspiracy theory. But once he was out of office, prominent scientists came out saying the hypothesis was legitimate. Political fights like this hurt our government’s ability to track reality. Having Congresspeople with an eye towards the long term in both parties, who could break party lines to support especially impactful policies, seems really desirable.

My plan

I just donated the max allowed, $2,900, to a candidate I thought was especially promising and well vetted by people I trust. Going forward, for my EA donations, I’m planning to prioritize 1) opportunities where I have special knowledge and 2) political campaigns of aligned congressional candidates. My current best source of knowledge is politically oriented EA friends and Guarding Against Pandemics. So I plan to mostly donate where they recommend and do some spot checking as I go. If you find these arguments compelling, I recommend you do the same! 

You can sign up via this form to be notified about campaign donation opportunities in the future.

51

20 comments, sorted by Click to highlight new comments since: Today at 8:32 AM
New Comment

I think the fact that contributions and contributors are publicly disclosed by campaigns at fixed intervals is another argument in favor of contributing to candidates (perhaps over other giving options) — particularly for EAs who may not be giving the max to a specific candidate. The number of contributions and/or contributors can often be used by political insiders/press as a signal for how well a campaign is doing, so for every new contribution/contributor, there is a small downstream effect of potentially influencing other people to also support that candidate. I'm not sure this dynamic necessarily exists for contributions to non-profits. In addition, the average contribution amount is also often used by political insiders/press as an indicator of how much "grassroots support" a campaign has.

I would also encourage EAs to contribute to non-federal candidates, i.e. state/county/municipal candidates and/or party candidates (such as precinct chair, county/national/state committee, etc). While these roles may be less directly relevant to x-risks, I think the probability of your contribution making an impact in a race is significantly higher. And candidates who win those seats can also eventually influence federal races in addition to creating state/county/municipal policy on things like pandemic prevention and through the budget allocation process.

Disclaimer: I have personally made campaign contributions and have worked on campaigns

Also important to note: U.S. green card holders / permanent residents can also make financial contributions to political campaigns, so this opportunity is not limited to U.S. citizens!

Also, anyone can volunteer for campaigns afaik!

Note that large funders such as SBF can and do support political candidates with large donations via PACs, which can advertise on behalf of a candidate but are not allowed to coordinate with them directly. But direct donations are probably substantially more cost-effective than PAC money because campaigns have more options on how to spend the money (door-knocking, events, etc not just ads) and it would look bad if a candidate was exclusively supported by PACs.

The optics concern makes sense to me, but I'm 90% certain PACs and Super PACs can and do spend on things that are not ads? Eg paid canvassers/phone bankers, polling, mailers, etc.

Additionally (and I'm not advocating for this), there seem to be many ways to get around the coordination ban, e.g.: https://www.nytimes.com/2020/02/18/us/politics/buttigieg-votevets-super-pac.html

They can do things other than ads. The real limitation is that by law candidates are entitled to the best unit rate on ads that a tv, cable, or radio station offers, and other political organizations are not. And the rates for other orgs can be quite a bit higher (the linked article cites other articles saying up to 6 times as much). Additionally, especially early in the race, how much money a candidate has raised directly for the campaign will be taken by other potential donors, volunteers, endorsers, the media, etc. as a sign of whether it is viable, and there are pretty solid reasons for this related to whether the candidate is actually putting in the work, so I don't think money a Super PAC plans to spend would be counted this way in their eyes. So if you like what a PAC or a Super PAC is doing, it's more cost-effective to look at the candidates they support and just give the money directly to the candidate. 

Ah, this is interesting about the ad rates-thanks for the link!

My understanding is there are different kinds of PACs. Guarding Against Pandemics has a PAC that cannot receive more than $5000 from individual donors, but can donate up to $5000 directly to campaigns.

Are there one or more reasons you didn't mention by name the candidate whose campaign to which you've contributed, Carrick Flynn, a participant in effective altruism who has a professional background in x-risk reduction, as opposed to only being someone who has privately been vetted by trusted peers?

I have more-or-less come to this same conclusion. As I mentioned in a reply below, Guarding Against Pandemics has a PAC that can receive up to $5000 from individual donors, and can, in turn, donate up to $5000 directly to campaigns. 

As the linked post explains, "Donations to the PAC would go towards supporting candidates who are champions for pandemic preparedness in Congress and in state and local offices. ", so, not necessarily EA aligned in other ways. They could also be from either political party.

This seems potentially pretty impactful, but probably more risky than donating directly to an EA aligned candidate. I am curious what others think about this, or if anyone has done an analysis or anything.

I don’t know much how campaign spending relates to election outcomes

This "uncertainty" is extremely relevant to how effective such donations are. I recommend attempting to determine how effective campaign contributions are before promoting this "donation opportunity." The political science literature suggests that the effect of donations in congressional elections is generally about zero.

There's pretty solid research on how to use money effectively in campaigns, and some low-quality, underexplored  research that suggests even better options than what the RCTs suggest. So if the lit says donations make no difference it's very likely that either campaigns are spending their money poorly, or it's just poor quality research. So you could simply donate the money to campaigns that plan to spend it effectively. 

 

I don't think donating willy nilly to non-federal candidates makes sense. If these are EAs or others who you expect to have a strong alignment on x risks, and who you could plausibly see being a candidate for Congress someday, then it makes sense, after you've exhausted donation opportunities for strong-on-x-risk current federal candidates.  Yes, your contribution has a higher chance of influencing the outcome of a smaller race, but so what? The stakes are just way, way smaller. Even if you were to say, target Massachusetts state legislative candidates in the hopes of banning gain of function research right in the heart of biotech, what does that get you? One year of delay while the researchers transfer to labs in California? And can the state even legally implement a ban when the federal government is funding the research? 

 

OP, do you have a source for the claim that banning gain of function research is a policy that nearly everyone working on long-term issues would like to see happen? I thought this was the consensus at one point, but there do seem to be potential benefits from it, so I'd like really like to see some harder numbers of what % of longtermists with issue are expertise support this view and how much confidence they hold in it. In fact I originally came to the forum tonight to look for something like that, then got distracted by this post. 

There's pretty solid research on how to use money effectively in campaigns, and some low-quality, underexplored research that suggests even better options than what the RCTs suggest. So if the lit says donations make no difference it's very likely that either campaigns are spending their money poorly, or it's just poor quality research. So you could simply donate the money to campaigns that plan to spend it effectively.

  • What are you thinking of? (The only thing that I've heard works to change minds is deep canvassing, and I'm not aware of campaigns doing this.)

you could simply donate the money to campaigns that plan to spend it effectively.

  • Can your name a couple?

It would not be correct to dismiss the political science consensus as bad research. And regardless of whether some spending is effective, the fact is that spending more doesn’t causally affect congressional candidates’ prospects.

+1 to @BlueFalcon's response below, and I would also add that:

  • There is indeed research that deep canvassing is effective, but the set of things that you can spend $ on to change people's minds ("persuasion") is only a very small subset of things that you can spend $ on in campaigns. For instance, spending $100 to ensure that a voter who already supports you turns out to the polls and votes does not change any minds but does net 1 vote.
  • There are thousands upon thousands of RCTs compiled over at Analyst Institute on how campaigns can effectively spend $. I shouldn't share anything too specific, but the research ranges on anything from messaging in a social media ad, timing of certain campaign expenditures, voter persuasion, etc.
  • Separately from all of the above: I think it is also true that from a donor's perspective, it is probably difficult to tell which campaigns are spending $ effectively. This is likely a combination of the following factors:
    • Most effective $ allocation will differ based on the candidate, party, demographic, district, etc.
    • Campaigns are (perhaps understandably) reluctant to discuss campaign strategy with basically anyone
    • There is no (publicly-shared) "impact evaluation" for campaigns
    • Many campaigns may in fact not be spending $ effectively because campaign staffing is extremely talent bottlenecked, and the Peter Principle probably holds strongly in many campaigns (esp smaller ones)
    • Campaign strategy shifts quickly

Love the Analyst Institute, which has done herculean amounts of work on figuring this stuff out.

 

Curious why you think there's an extreme talent bottleneck  for campaign staffing. My impression is they may be hiring the wrong people (i.e. too many people with lots of "experience" but not enough with experience on a modern campaign), but I suspect most decently well-funded campaigns could get the kind of people they wanted if they in fact wanted the right people. 

Hmm,  I should probably be more specific in defining what I mean by "talent bottleneck in campaigns." There is probably less of a bottleneck for large campaigns at the presidential/gubernatorial/senate level; I would estimate that on smaller races from House and non-gubernatorial state-level races downward, there is:

  • A shortage of campaign staffers who have a skillset in campaign management / strategy—I would argue that campaign management / strategy is sufficiently distinct from non-profit / industry / other organizational management / strategy to warrant categorizing as a bottleneck
  • A shortage of staffers who are highly-skilled in field / organizing. There seem to be a large number of people who are moderately-skilled at this, but a shortage of people who are highly-skilled (weakly-held opinion derived from my personal experience, so I could be wrong about this one)
  • There is probably not a shortage of people who have the skills for data or digital work, but there is a severe lack of funding + pay is not competitive for these roles in smaller races. There may be a shortage of people with experience working with campaign-specific tools such as VAN or Action Network, but I don't think this is a bottleneck necessarily as these can be easily picked up if you have certain technical skills

To your point above, there is probably also an information problem in the campaign staffing job market in the sense that:

  • Prior work history on campaigns seems to be an extremely poor indicator for whether someone is actually good at working on a campaign
  • There are no good credential / educational indicators that would predict whether someone would be good at a campaign role; this may have changed somewhat with some of the orgs listed below but does lead to things like a severe bias toward hiring from elite schools in Democratic circles (which I think is very bad)
  • I have a weakly-held belief that most campaign management does not actually know how to hire, so even if they did have funding, they would probably not ask the right questions or do the right outreach to get talented campaign staffers hired

Some of these issues do seem to have gotten better in recent years with the advent of organizations/programs like the NDTC, Arena PAC, Movement School, various apprenticeship/bootcamp programs, etc. There are probably also conservative / Republican analogues to these programs, but I have no clue what they are as I don't work on that side of the aisle.

Hmmm....what specific skills are the people getting hired in management and field roles missing?  If you can break that down further maybe it's possible to screen for those specific skills. And digging a bit further into this, how do you know the management problem is with the CMs and not the candidates? At the lower levels, you have this weird situation where you get the top job (candidate) by just showing up, but then there's actually a selection process for the second in command, run by the person who got their job by just showing up. 

I go back and forth on my opinion about hiring from elite schools. IME the schools are quite good at skimming the cream, so you will generally be stuck with less talented people if you don't hire from there. OTOH  the culture of elite schools is not the culture of most places in America and staffers' failure to understand that seems to be a real problem. So idk what the solution is. Hire from elite schools, but only grads who grew up in the state you want them to work in, or a similar state? Hire from the top 10% state schools? 

But maybe the inability to get elite school grads to run competent campaigns does point to the absence of skilled campaign managers. After all, these are people who have been quite good at responding to incentives from an early age, so why are they not sufficiently incentivized to understand local values? 

I'd probably have to think harder about  breaking down the specific skillsets; re: the comments above, hiring seems to be one of the skills that would fall into this subset. It would likely be slightly different for different management roles. 

I agree with the weirdness in smaller races, but I think that this may be more of a culture issue than a talent issue—candidates should just learn to rein in their egos, step back, and realize that running a campaign is not a skillset that they have (and that they should let their staff handle it). I've worked with candidates who are very good about this, and it makes life much easier on a campaign.

Re: elite schools, I think the part of elite grads being disconnected from the reality of what happens in actual communities is probably true. But a bigger question may be, why hire at schools at all? I'm not sure that the set of skills one needs to be good at campaign roles is even weakly correlated to academic performance / admission to a top US university; people just seem to default to this as a proxy for hiring, and I don't think it's a good proxy.

Political science consensus? Setting aside the generally poor quality of social science research, as noted in the recent replication crisis, Green and Gerber wrote a whole book on what works. See https://www.amazon.com/Get-Out-Vote-Increase-Turnout/dp/0815736932/ref=sr_1_1?crid=1OHCY3DHNYA8S&keywords=get+out+the+vote&qid=1651864901&sprefix=get+out+the+%2Caps%2C87&sr=8-1.

Canvassing is the most cost-effective thing anyone's run RCTs on, but other things (e.g. phone calls) do definitely move votes. 

A lot of mass media stuff no one has really run good experiments, and political scientists seem to have a bias against doing the experiments because canvassing is more of a feel good story than "yes you can influence people's decisions by impersonal corporate-feeling means", but the little bit of research that has been done suggests print, radio, and tv are all more cost-effective than canvassing. 

 

Mailers and robocalls are generally ineffective but that's a far cry from "spending money doesn't work". 

Zach, can you provide a source that it is "about zero"?

Hmm, no. I retract my "about zero" and "political science consensus" claims until learning more.