Strategy Fellow at Open Philanthropy, working to help identify new cause areas within the Global Health and Wellbeing portfolio. Based in Brussels, grown in the UK.
Giving What We Can member since 2011. Previously earning to give as a strategy consultant.
Not the Chris Smith who used to work at GiveWell.
Tweets on global health, statistics, economics, feminism, and effective altruism at @chris_topian
Yes, we do consider the benefits of alcohol, including that many people enjoy it.
James Snowden put together a short document discussing this when he made the largest current Open Philanthropy alcohol grant in 2021 (the grant was recommended by GiveWell but funded by Open Philanthropy; any extension / renewal will sit within Open Philanthropy). At the time GiveWell / James applied a 10% reduction to the (implicitly net) burden of alcohol harm on this basis.
I’m reviewing this issue in greater detail now.
GiveWell / Open Philanthropy have funded CHAI.
I think there’s value in some organizational diversity / having multiple shots on goal, and I’m excited to see what comes out of both CHAI’s ongoing work and efforts by Charity Entrepreneurship and others to create effective non-profits.
Thanks for writing this Nick, I’m sympathetic and strongly upvoted (to declare a small COI, I work at Open Philanthropy). I will add two points which I don’t see as conflicting with your post but which hopefully complement it.
Firstly, if you’re reading this post you probably have “EA resources”.
You can donate your own money to organisations that you want to. While you can choose to donate to e.g. a CEA managed EA Fund (e.g. EA Infrastructure Fund, EA Animal Welfare Fund) or to a GiveWell managed fund, you can equally choose to donate ~wherever you want, including to individual charitable organisations. Extremely few of us have available financial resources within two orders of magnitude of Open Philanthropy’s core donors, but many of us are globally and historically rich.
It’s more difficult to do well, but many of you can donate your time, or partially donate your time through e.g. taking a lower salary to do direct work on pressing and important problems than your market value in another domain.
Secondly, and more personally, I just don’t think of most of my available resources as being deservedly “mine”. However badly my life goes, save civilizational catastrophe, I will not die because of lack of access to low-cost medication. I will not be slaughtered painfully and then processed into a chicken nugget, or a fish cake, or shrimp paste. My relative wealth is not deserved, and justice calls me to use it well to make the world better.
Warren Buffet’s giving pledge letter explores a similar theme:
My wealth has come from a combination of living in America, some lucky genes, and compound interest. Both my children and I won what I call the ovarian lottery. (For starters, the odds against my 1930 birth taking place in the U.S. were at least 30 to 1. My being male and white also removed huge obstacles that a majority of Americans then faced.)
My luck was accentuated by my living in a market system that sometimes produces distorted results, though overall it serves our country well. I’ve worked in an economy that rewards someone who saves the lives of others on a battlefield with a medal, rewards a great teacher with thank-you notes from parents, but rewards those who can detect the mispricing of securities with sums reaching into the billions. In short, fate’s distribution of long straws is wildly capricious.
The reaction of my family and me to our extraordinary good fortune is not guilt, but rather gratitude.Were we to use more than 1% of my claim checks on ourselves, neither our happiness nor our well-being would be enhanced. In contrast, that remaining 99% can have a huge effect on the health and welfare of others.
Small suggestion—could you include some text on the front page about who you think the survey is for (e.g. is it everyone who self-identifies with the term effective altruist? anyone who considers themselves part of the EA community? someone who has read a book / listened to a podcast about effective giving / longtermism / farmed or wild animal welfare?).
I appreciate that the sampling frame here is extremely difficult and I’m supportive of trying to survey ~everyone of relevance, but the way it’s set up now it’s not clear to me who you’re trying to reach. I can imagine people who you might want to reach and fill the survey out not doing so because of how the landing page is set up. I’d push for an inclusive framing of who you’re trying to include. The current page assumes that the reader knows what the “EA survey” is—which is pretty ingroup / assumes a lot.
[Context—I managed the Cause Exploration Prizes]
Thank you Gavriel for taking the time to write this out and thank you again for your original submission on ways that philanthropic funders can help address indoor air quality, which I encourage others to check out. I’m really sorry to hear you felt burnt out after completing the entry.
Although essay prizes and contests are quite prevalent in the EA community, this was very much an experiment for global health and wellbeing cause prioritization team at Open Phil. A major objective of the lower value prizes and participation awards was to enable people to feel able to submit ideas that were good / on their mind, but not fully polished. Personally I think we were reasonably successful with that—but if we run anything similar again, we’ll definitely consider a smoother gradient and tweaking other parameters.
Yes—this fits within our GHW portfolio. From the FAQ page:
Can I write about non-human animals?
Yes. Open Philanthropy is a major funder of work to improve farm animal welfare. If you want to write about a potential new cause area where the primary beneficiaries are non-human animals, please use the open prompt.
Hi jserv! I’ll aim to say a bit more about the nuts and bolts of the process in an update before the end of the year, but prize selection was dominated by blind, independent review. I’m following up with you privately on tuberculosis.
The basic idea is that we mostly pay for drugs based on volume (e.g. if a manufacturer charges $50 per insulin vial, they want to sell lots of vials), but that mechanism is inappropriate for novel antibiotics, since there are major societal benefits to not over using new antibiotics but retaining them as last resort drugs. This means that it is ~impossible for developers of novel antibiotics to be economically viable. A subscription model de-links payment and volume—the provider is paid for having created and made the drug available, irrespective of volume used.
There’s a nice CGD paper discussing some of the specifics of the main US proposal on pull incentives / subscription models for novel antibiotics. The website AMR Solutions has broader coverage, including of the UK’s similar model.
Hi Nick, thanks for engaging.
Part of the rationale for the Regranting Challenge was that it might be possible to identify funders who already have the infrastructure to deploy funding effectively at scale, rather than creating duplicative philanthropic infrastructure. In 2021, Open Philanthropy recommended ~$400M of grants with an average staff number of ~40 (so ~$10M / FTE). To compare, the Gates Foundation granted $6.7B with a staff of 1,736 (so <$4M per staff person); Wellcome Trust moved £1.23B with a staff of ~800 (so <$2M per staff person).* Now Open Philanthropy is scaling up internally (see jobs page!), and we pay others to do research on our behalf, but these things take time so we wanted to experiment with this approach of working through other funders.
Scale / existing infrastructure isn’t the only benefit though—sometimes other funders have advantages that we don’t.
I think a good example of this is the award we made to the Eleanor Crook Foundation. Transparently, they know more about malnutrition than Open Philanthropy does, and they have relationships with researchers and decision makers that would take a long time for us to develop internally. They can make grants here that we would struggle to make as effectively, even though we have existing grantmaking in both development research and aid policy. As Alexander tweeted, every funder that cares about X having its own program X can’t be the best approach.
On the specific DIV example, we describe some of the rationale for this on the minisite:
DIV makes four types of grants:
Stage 1 (Pilot) grants support initial small-scale implementation of new ideas in low- and middle-income countries to test user demand, feasibility, impact, and financial viability (up to $200,000).
Stage 2 (Test and Position for Scale) grants support rigorous testing to determine impact or market viability (up to $1,500,000).
Stage 3 (Transition to Scale) grants transition proven approaches to widespread scale in new contexts or new geographies (up to $15,000,000).
Finally, DIV makes Evidence Generation grants to support research and evaluation of widely-used development approaches that are not necessarily innovative, but lack sufficient evidence of impact and cost-effectiveness (up to $1,500,000).
With $45,000,000 in Regranting Challenge funding, DIV will expand their grantmaking through Stages 1, 2, and 3 grants and launch a new “Stage 4” initiative. Stage 4 grants will help some of the most promising programs in DIV’s portfolio scale up through partnerships with Missions and Bureaus across USAID. They will provide partial match funding for other USAID spending, with the aim of expanding the use of innovations that have been rigorously demonstrated to cost-effectively improve development outcomes.
A key reason for working through DIV rather than trying to do this work directly is that they sit within a much larger pool of resources than we do, much of which is spent without the same focus on impact and evidence that Open Philanthropy, effective altruists, and DIV themselves bring to the table. To be clear—this might not work! But often you have to try things out to see.
*These figures are rough; I’ve not tried to adjust for e.g. part-time of definitional differences. I’d note additionally that Open Philanthropy makes a lot of very time intensive small grants (e.g. to individuals) mostly within the longtermism portfolio.
(I don’t lead on the air quality work, so be more careful with this comment that the others that I’ve left here).
India wasn’t picked as an example to illustrate the importance and neglectedness of air quality work. Rather, India has been the dominant setting for Open Philanthropy’s air quality work to date—it even has its own updated web page. You can read more about why Open Philanthropy launched the work on South Asian air quality here and here. Santosh Harish, the Program Officer who leads that work, recently gave an excellent interview to the 80,000 hours podcast—transcript and recording here.
I agree domestic financing complicates relative neglectedness—the effort here was to be as consistent as reasonably possible between the risk factors. Neglectedness comparisons are very tricky to nail down in general (e.g. how to attribute non-specific health systems spending across both causes and risks, whether to include treatment for linked health conditions like lung cancer or cirrhosis, how to think about relatively ineffective uses of money like e.g. biofuel subsidies for climate change, or a more relevant example here would be smog towers for air quality). One nod to the uncertainty of both measurement and scope here is the use of ranges; but yeah, we’re trading off a bunch of different considerations here.
There’s a lot of internal research stress testing the IHME burden estimates for lead and air quality, and some on alcohol—I’m doing more on alcohol specifically at the moment. Here we’re pointing to the IHME GBD study for several reasons: it’s widely recognized, easy to interact with, has a largely consistent / common methodology between different causes of death and disease, and importantly doesn’t allow for deaths to be attributed to more than one cause. This works well for problems where the Open Philanthropy way of conceptualizing the problem (e.g. malaria, lead exposure) matches a GBD cause (e.g. malaria) or risk (e.g. lead exposure). This doesn’t mean we uncritically use the GBD in all of our own decision making—but this set of reasons make it very helpful to refer to when communicating externally. It might be that we publish more of our internal research on this in the future, but honestly it’s a serious time investment and I don’t want to over promise.
DALYs sit behind the framework but can be understandably offputting for many audiences. The BOTECs / grant decisions are in line with our usual GHW cause prioritization framework of valuing increases in healthy life and log-income.
Thank you to the survey team for completing what is an easy-to-underestimate volume of work. Thank you also to the many who completed this survey, helping us to both understand different EA communities better and to improve this process of learning about ourselves as a wider group in future years.
I have designed and analysed several consumer surveys professionally as part of my job as a strategy consultant.
There is already a discussion of sample bias so I will leave those issues alone in this post and focus on three simple suggestions to make the process easier and more reliable for when this valuable exercise is repeated next year.
Firstly, we should use commercial software to operate the survey rather than trying to build something ourselves. These are both less effort and more reliable. For example, SurveyMonkey could have done everything this survey does for about £300. I’m happy to pay that myself next year to avoid some of the data quality issues.
Secondly, we should use live data validation to improve data collection, data integrity and ease of analysis. SurveyMonkey or other tools can help John to fill in his age in the right box. It could refuse to believe the 7 year old, and suggest that they have another go at entering their age. It could also be valuable to do some respondent validation by asking people to answer a question with a given answer, removing any random clickers or poor quality respondents who are speeding through (eg “Please enter the number ‘2’ in letters into the textbox to prove you are not a robot. For example, the number ‘1’ in letters is ‘one’”)
Thirdly, we should do more testing by trying out draft versions with respondents who have not written the survey. It is very, very hard to estimate how people are going to read a particular question, or which options should be included in multiple choice questions. Within my firm, it is typical for an entire project team to run through a survey several times before sending it out to the public. Part of the value here is that most team members were not closely involved in writing the survey, and so won’t necessarily be reading it in the way the author expected them to read it. I would suggest you want to try any version of the survey out with a large group (at least twenty) of different people who might answer it, to catch the interpretations of questions which different groups might have. Does the EA affiliation filter work as hoped for? Are there important charities which we should include in the prompt list? It does not seem unreasonable to pilot and redraft a few times with a diverse group of willing volunteers before releasing generally.
The analysis throws up several interesting conclusions, and I have learned a lot by reading through it. The main shocks are: the relatively low levels of donations in $ terms by many self-identified EAs, the relatively low proportion of EAs identifying chapters/local groups as a reason for joining or identifying with the community and, (for me) the encouragingly high proportion of respondents who are vegetarian or vegan.
I’m going to set aside some time in May to go through the data in a ‘consulting’ sort of way to see if that approach throws up anything interesting or different to others and will circulate with the survey team before publishing here.
Thanks for sharing Justin!
I’d recommend anyone who’s ever worried about our ability to make global progress against big problems read the annual updates from UNAIDS in the late 1990s/early 2000s and compare to more recent updates.
Drawing on your lessons above, what are the PEPFARs we should be pushing for today?
This is a really helpful post—thank you! It does blow my mind slightly that this isn’t more broadly practiced, if the argument holds, but I think it holds!
I don’t know enough about the market for academic papers, but I wonder if you’d be interested in writing this up for a more academic audience? You could look at some set of recent RCTs and estimate the potential savings (or, more ambitiously, the increase in power and associated improvement in detecting results)
Given that the argument is statistical rather than practical in any way that is specific to economics or development, do you know if this happens in biomedicine? Many trials often involve pitting newer, more expensive interventions against an existing standard of care.
I’m going to reproduce a comment I wrote at the time the 2014 results were released in order to have them on the agenda for the call later on. I remain convinced that each of these three practical suggestions is relatively low effort and will make the survey process easier, the data more reliable and any resulting conclusions more credible:
Firstly, we should use commercial software to operate the survey rather than trying to build something ourselves. These are both less effort and more reliable. For example, SurveyMonkey could have done everything this survey does for about £300. I’m happy to pay that myself next year to avoid some of the data quality issues.
Secondly, we should use live data validation to improve data collection, data integrity and ease of analysis. SurveyMonkey or other tools can help John to fill in his age in the right box. It could refuse to believe the 7 year old, and suggest that they have another go at entering their age. It could also be valuable to do some respondent validation by asking people to answer a question with a given answer, removing any random clickers or poor quality respondents who are speeding through (eg “Please enter the number ‘2’ in letters into the textbox to prove you are not a robot. For example, the number ‘1’ in letters is ‘one’”)
Thirdly, we should do more testing by trying out draft versions with respondents who have not written the survey. It is very, very hard to estimate how people are going to read a particular question, or which options should be included in multiple choice questions. Within my firm, it is typical for an entire project team to run through a survey several times before sending it out to the public. Part of the value here is that most team members were not closely involved in writing the survey, and so won’t necessarily be reading it in the way the author expected them to read it. I would suggest you want to try any version of the survey out with a large group (at least twenty) of different people who might answer it, to catch the interpretations of questions which different groups might have. Does the EA affiliation filter work as hoped for? Are there important charities which we should include in the prompt list? It does not seem unreasonable to pilot and redraft a few times with a diverse group of willing volunteers before releasing generally.
I am grateful for this post and think it demonstrates bravery that Rohit didn’t need to show. He’s a thoughtful, accomplished professional who has approximately no personal incentive to writing this out.
I hope readers who wish for a healthy community around the ideas of effective altruism, and who want thoughtful engagement from people exploring similar questions that effective altruists consider important in good faith, reflect on the damage that the discourse of the past few days (and the incident that kicked it off in terms of Bostrom’s poor statement on historical failings) has caused to the mission to make the world a better and safer place.