RW: You said that you thought that most of the increase in cost-effectiveness - or your hopes of finding charities that are more cost-effective - came from different sorts of causes, such as scientific research, catastrophic risk and migration. What are some of the top priorities at the moment?
EH: They are areas about which we currently know very little: the broad area of funding scientific research and the broad area of advocacy. An example of the latter would be lobbying for increased foreign aid or better allocated foreign aid or advocating for something related to immigration regulations across countries. These are areas that we find appealing. So far, we’ve published a page on climate change and a page on asteroids - both of those would fall into the catastrophic risk category. We’ve also published a page on seasonal migration in low-income countries. Basically, we see these big areas to look into. Broadly, we’re open to pretty-much anything that could have a significant impact on humanitarian outcomes. We’re trying to come up with a systematic process to briefly investigate different areas, share what we’ve learned, and then be in a position to decide which programs deserve increased attention from GiveWell. The area we’ve actually spent the most time on is meta-research - primarily medical meta-research. This means trying to improve the replicability and reproducibility of medical trials, which to me, somewhat surprisingly, is a field that could use some support.
RW: So I imagine in a lot of these other causes that you’re hoping to investigate and get recommendations from in the future, it’s going to be harder to get firm evidence of impact in the same way that was possible in global health. Are you going to, to some extent, lower the standards or be willing to accept greater risk in those recommendations?
EH: It’s definitely going to be totally different. I think a nice quality of the research and recommendations is that they’re about as quantifiable and measurable as I think charity can be, and even then I think there’s a large aspect of the conclusions that we’ve reached that are not based entirely on quantifiable information. As we move onto new areas, they’re obviously much less measurable, quantifiable, objective, and so therefore our standards are going to remain high. Our standards of critical analysis, transparency, open discussion will be as high as ever. In some ways, I think we will need to have an even higher focus on our critical analysis, because it’s so much easier to go down a bad path when you don’t have that quantifiable outcome or that very measurable bottom-line to hold yourself accountable to. It is risky and it is difficult, but we want to do this - we feel the need to do it - because if they could be more impactful than the organisations we’ve found today, then we have to look at them. GiveWell is not about minimising risk or finding high-probability-of-impact charities. What we broadly care about is having the most humanitarian impact per dollar, and if that means having more risk, there is absolutely no reason that we would stay away from it, I mean we intend to be very bold.
RW: GiveWell started GiveWell Labs to try to get new and exciting projects off the ground – how's that going? What kinds of things are you funding, if any, and how much money is in the pot?
EH: We're open to getting new organizations off the ground, providing partnership to existing organizations, open to writing a blank check outside of global health and development that doesn't have the same type of evidence. All that is included in GiveWell Labs. The biggest focus that we see now for GiveWell Labs is figuring out which causes we want to be involved in. Do I want to focus in on global health? Is that the place GW Labs can have the most impact? Or should it be a particular causes such as advocacy? That's the project that we're undertaking. This is all at very preliminary stages, so what is out there so far are the shorter investigations of new causes that we've published, some conversations we've had with various experts in different areas, and then our research meeting audio and transcripts. But broadly GiveWell Labs today is sort of where GiveWell itself was in the fall of 2007, beginning of 2008. Preliminary stage of trying to figure out how we should go about doing the work.
RW: What difficulties with giving unrestricted funding to charities were you trying to address by starting GiveWell Labs? Do you want to have more control over the projects or do things that no existing organization was working on?
EH: There are a few different aspects of GiveWell vs. GiveWell Labs. GiveWell direct aid is what the current GiveWell was up until the creation of GiveWell Labs. We were looking for proven cost-effective organizations that can take in more money and do more good. What we were looking for was programs that had strong evidence of effectiveness, organizations that could take in an arbitrary amount of money (we recommend them, they could get $250,000 or $5 million) and they would accomplish more good. When we moved to GiveWell Labs we want to be able to do something else, consider causes that don't have the same evidence of effectiveness. There's no way one could make an argument for the cause of meta-research using the same criteria that we have used for GiveWell direct aid. Bed nets are a very different animal than funding literature reviews or reproducing existing studies. We want to look into those areas, this is something different, therefore we want to call it something different and see it as a different activity. There is this other change that goes from providing unrestricted to project funding. I think the biggest driver is that as you move into these other areas, the most common way organizations work is through the provision of project-related funding - the unrestricted vs. project distinction is not major. If we have a particular focus and we find relevant organizations, then sure, unrestricted funding works. On the other hand, we might want to play a more active role in the funding process. Major philanthropists could say that they want to create these organizations or that they want a set of organizations to focus on these particular issues that they may not have otherwise focused on. That is also a diversion from what we've done in the past, from being more passive and funding existing organizations to do the things they already want to do, to being more of a driving force behind the activities.
RW: On a different note, I wrote a blog post suggesting that it should be possible to beat the cost-effectiveness of GW's top recommended organizations by instead giving to an organization that would fundraise for those charities. It seems like each dollar spent on fundraising raises $4+ for whichever organization - even using fairly mundane fundraising methods like having people speak to strangers on the street. Do you think this argument makes sense? If so, does it have any consequences for what our members or anyone else should be doing?
EH: I wonder what led you to the conclusion that a dollar on fundraising creates more than an additional dollar of donations? If it were true, charities should spend significantly more on fundraising, that would make them successful. The major question is, do you feel that is the case?
RW: My impressions are that it’s something that is generally accepted among fundraisers that they can pay people to do this and raise more money than what it costs them. Your main doubt is around the empirics of that? If there were strong evidence, would it potentially change your recommendations, or attract more attention from you?
EH: It could. I'm pretty skeptical that that's the case. A lot of organizations out there could easily increase the amount of money they spent on fundraising. I don't have specific knowledge, but Doctors Without Borders is a gigantic organization. They're credited with potentially creating direct-mail fundraising. Very successful. They also have a relatively low fundraising ratio. They could increase the fundraising ratio – why don't they? If it were as simple as that case make out, you should see charities behaving differently.
RW: That's the big question, why wouldn't they just spend more money on fundraising? Donors are really averse to organizations with a large overhead - this is a large obsession. If a charity was spending too much on fundraising, that would deter people from giving in the first place. Do you think that's a possible explanation?
EH: I'm skeptical that it's the main driver of what's going on here. Fundraising is hard. At GiveWell, I don't feel like we could spend an additional dollar and raise two additional dollars for our top charities. We just don't see the curve of what happens as we scale up money spent on fundraising driving more money to our top charities. I think there's a lot of charities out there – why don't they spend more? Or you're making an argument for why they don't. I think there's all these organizations out there that are incredibly successful, professional, talented at raising money, and they give their fundraising ratios a relatively low amount. If it were as simple as this argument implies, it doesn't make sense that they would do this.
RW: What if you were a small charity like AMF? If you're a large organization, everybody already knows you, it's possible you could speak or send something to everyone who would be likely to donate. But there are many people out there that haven't heard of AMF – do you think there might be more low-hanging fruit yet?
EH: Maybe. They're getting letters from MSF, from Nothing But Nets, I think it's a more competitive, more challenging space than is implied. I think the bigger picture is, it doesn't quite sync up that this should be obvious, and therefore I would assess this case on the basis of a specific organization demonstrating that it would be able to do this. It's a competitive field, there are incentives for people to go into this and be successful, so all of those are reasons that the theoretical case is not particularly compelling to me. If someone said, “I am able to do this,” and address all the questions, particularly on whether they were just reallocating funds, which is challenging to determine, that would be interesting. But it doesn't strike me as one of the more promising endeavors.
Look out for the last part of the interview on this blog soon.
Image credit: NASA/ESA and D. Jewitt (UCLA)