# The GiveWell Blog

Following up on the inaugural open thread, we wanted to have another one.

Our goal is to give blog readers an opportunity to publicly raise comments or questions about GiveWell or related topics (in the comments section below). As always, you’re also welcome to email us at info@givewell.org if there’s feedback or questions you’d prefer to discuss privately. We’ll try to respond promptly to questions or comments.

• Joseph Porter on April 16, 2015 at 8:37 pm said:

I appreciated and have benefited from GiveWell’s advice on how to maximize the impact of my charitable contributions. Any thoughts on maximizing volunteer hour effectiveness in addition to donated dollar effectiveness?

Many people strive to be generous both with their money and with their time. Does GiveWell have suggestions for those people on how best to spend their time?

• Does GiveWell (or: do individual GiveWell employees) have opinions, even if not robustly evidence-backed opinions, about the effectiveness of large big-name charities like, say, Oxfam?

Closely related: To what extent are small single-issue charities like AMF favoured in GiveWell’s evaluations because they’re *actually better* (having, e.g., found some juicy low-hanging fruit to pick), and to what extent just because they’re easier to evaluate?

(I mention Oxfam specifically because they seem to be the big-name charity most frequently advocated in EA circles; e.g., they’re one of the ones listed at thelifeyoucansave.org. But of course there are lots of others. I am thinking specifically of big-name charities that try to help the quality of life of poor people in poor countries; of course it’s clear enough why the Cute Sick Puppy Foundation might be ineffective but popular.)

• Colin Rust on April 18, 2015 at 11:29 am said:

For better or for worse, a lot of people look at Charity Navigator when they hear about a charity. Unfortunately, GiveWell and most of its recommended charities aren’t even covered by Charity Navigator (even when eligible). They have a process by which you can vote to have a charity added. For instance, here is the link to vote for AMF. To vote for GiveWell, note it’s official name is Clear Fund.

(I’m only posting one link since I’ve found in the past comments with multiple links may not go through.)

• Colin Rust on April 18, 2015 at 11:40 am said:

Joseph, most opportunities to volunteer are local, so it’s probably hard to come up with general advice. Of course there are things you can do online; see my comment above for one.

Another option is to earn money with your extra time and then donate that to a top charity. Depending on your circumstances, that may theoretically the most efficient, though it probably doesn’t feel as fulfilling.

• Any comments on New York Times article, “Meant to Keep Malaria Out, Mosquito Nets Are Used to Haul Fish In”?

The article is available at:

http://www.nytimes.com/2015/01/25/world/africa/mosquito-nets-for-malaria-spawn-new-epidemic-overfishing.html?_r=0

• Sorry, just saw GiveWell’s response to the bed-nets-for-fishing problem here:

https://blog.givewell.org/2015/02/05/putting-the-problem-of-bed-nets-used-for-fishing-in-perspective/

I earlier had looked at GiveWell’s general page on malaria nets and did not see nets for fishing as one of the drawbacks discussed.

• Alexander on April 20, 2015 at 2:25 pm said:

Joseph – unfortunately, we don’t have great suggestions about where to volunteer. We think it’s a much harder problem to address because an individual’s skills and interests are much less fungible than money, so the appropriate recommendations will likely depend a lot more on the specific context. We’ve written about volunteering before here and here.

G – broadly, we think it’s a combination of the two issues you suggest: organizations that effectively carry out individual cost-effective interventions are easier for us to assess and are also likely to be more cost-effective than organizations that conduct a broader variety of interventions (and don’t have a stringent cost-effectiveness filter). But it’s hard or impossible to provide conclusive evidence on the substantive point because we’ve never been able to get the level of visibility into the larger organizations that we do into the smaller single-issue ones. We’ve written about these kinds of organizations before here.

WO – thanks for the suggestion. We normally only update our intervention reports once a year, but I’ll add this to a list of updates to consider implementing.

• Hi, I wondered if GiveWell had any recommendations for environmental charities? I know that you prefer to focus on organisations the results of whose work can be measured accurately, but given that climate change will affect the planet on a massive scale, and the poorest people will be affected the worst, it would seem like an important issue to address.

• Thanks! The article you linked to says “We will also likely be investigating Oxfam”; I take it that never happened?

• Josh Jacobson on April 20, 2015 at 5:30 pm said:

I’ll apologize for the length of this in advance. That said, I think you’ll find each of these questions to be well-informed and address issues that many find interesting.

1. How does GiveWell feel about impact investing? Even though current efforts (such as Acumen) don’t meet its criteria, can GiveWell envision some form of new organization, dedicated to transparency and results, involved in investing that it would recommend alongside/in addition to the charities?

Would that organization have to live up to charity recommendation standards, or would GiveWell have a way to discount the expected results given that a financial return is provided?

2. GiveWell has said that it is partially happy not to focus on outreach given that other organizations do so instead. Are there any concerns that the organizations performing outreach generally recommend a far broader (such as x-risk, animal causes) range of donations than GiveWell? Would GiveWell consider taking a more active role to encourage more of these donations go to poverty-related causes?

3. If someone would like to participate in encouraging donations to GiveWell/its recommended charities, how would you recommend they best prepare for defending cause selection (health interventions v. any other, especially other foreign causes)?

The discussion on GiveWell’s blog seems much weaker than it is of why particular organizations are chosen.

4. Has GiveWell, or the leadership, experimented with different ways of describing the organization (online or in person) in order to determine relative efficacy? What are the lessons from these?

One example would be to describe itself as a way to fix the incentive structure in charities (effectiveness v. survival through feel-good marketing) rather than purely as an evaluator.

5. Would GiveWell consider strengthening its association with effective altruism, including the term in the about page or within the header/footer? Why or why not?

6. What one thing would GiveWell most want to see from the effective altruism community, other EA organizations, or a new EA organization?

7. Would GiveWell ever consider holding or publishing a training course? It seems as though there are many in the EA community who feel they could better defend EA causes if they were also trained/has tried to perform an example of the original GiveWell-level research themselves.

If not, could a particularly determined EA shadow a GiveWell researcher, or volunteer?

8. Has GiveWell experienced backlash for charities it has not recommended (from the organizations or the public)? Has there ever been any concern (or research into) the possibility that finding many organizations not worthy of donations decreases overall giving?

9. Holden granted an interview for one of the Money for Good studies. Has GiveWell engaged with, or used these studies in any way since?

10. What provides GiveWell the most credibility? Are there certain economists, philanthropists, or development practitioners who have endorsed its results and whose endorsement is particularly beneficial?

11. Can you alleviate our fears that with the OPP splitting off, GiveWell will still have the resources and ambition to potentially expand and grow? To the public, the limited number of nonprofits it covers is concerning, so many more shallow investigations/write-ups, especially of the most prominent organizations, would be helpful.

12. GiveWell’s budget reflects a very high amount toward a web redesign. What can we expect to see from this? Is it already in progress?

• Alexander on April 20, 2015 at 10:36 pm said:

Paul – thanks for the question. We’ve done some preliminary exploration around climate change in the context of the Open Philanthropy Project, but we haven’t prioritized further work in the area as a whole because we see it as fairly crowded relative to other comparably important causes. That said, in the context of our work on global catastrophic risks, we are interested in research on geoengineering, which we see as relatively neglected, but we don’t have public recommendations in the area at this point, sorry!

G – yes, that’s right.

• Holden on April 24, 2015 at 12:15 am said:

Josh, thanks for the questions. Quick responses:

1. We generally focus on giving rather than for-profit investing, and if we did encounter a strong opportunity for impact investing, we’d largely try to evaluate the good accomplished by the “expected losses” or other approximation to “money foregone” as though the latter were a gift.

2. We’re not concerned that other organizations recommend a broad range of things, and indeed we see significant value in causes other than direct aid targeting global poverty. The presence of outreach-focused organizations is only one relatively small factor in how much we prioritize outreach.

3. We’ve addressed many cause-selection-related questions at one point or another, but haven’t done a great job organizing what we have. We’re hoping to improve on that with our upcoming website redesign, which should be launching in the next month or two. We do have a summary of the case for developing-world-focused aid specifically: https://www.givewell.org/giving101/Your-dollar-goes-further-overseas

4. We have informally experimented with different ways of describing ourselves. Doing so formally has not been a priority. To some extent, different messages are best for different audiences, but the mission statement and other content on our current About page represent our best general description available.

5. At this time, we don’t plan the sorts of changes you refer to, because we feel that many people who are enthusiastic members of GiveWell’s audience would be put off by some of the messaging of other effective altruist groups. Our primary goal is to serve our audience and leave them feeling enthusiastic, so we don’t wish to connect ourselves in the ways you’ve outlined to a set of messages that is not core to our work. That said, we are open about our enthusiasm for effective altruism, and often bring it up informally.

6. This isn’t a question where we have a clear consensus answer, but we are aiming to make a post sometime in the next few months that goes through some of the organizations and efforts that we’d particularly like to see someone start.

7. This could be a possibility down the line but isn’t a priority for us at the moment.

8. We’ve had some pushback along these lines, but so far nothing sustained or prominent enough to have caused a major problem for us.

9. We wrote about the first Money for Good study here. We thought it was an interesting study and were glad to have more information about the potential size (both absolute and relative) of demand for research-based giving.

10. We list endorsements here, and other sources of credibilty here. We aren’t sure what the biggest contributors to our credibility are, but suspect (based on our conversations with major donors) that a lot of it comes down to having good answers to questions we get about our research (in conversation, on our blog, on conference calls) and having strong word-of-mouth-based reputation.

11. We have made a deliberate choice to recommend only a few charities, and this isn’t something we plan to change. GiveWell grew last year in the ways we consider most relevant, and plans to continue to do so.

12. We expect to launch new websites for both GiveWell and the Open Philanthropy Project in the next 1-2 months.

• Colin Rust on April 25, 2015 at 10:08 am said:

Holden, following your link at #3 above, it looks good in general, but in this sentence:

A child may die of malaria for lack of a $10 bednet, or of diarrhea for lack of a 5-cent packet of nutrients. I think the$10 is quite high. Even if you consider it as a fully loaded number, including administrative, distribution and monitoring costs, $10 is high. I think most people would read that though more narrowly as just the cost of the net which as you know is around$3.

• Colin Rust on April 25, 2015 at 11:34 am said:

On the Charity Navigator (CN) issue, it looks like in the case of the Against Malaria Foundation (AMF), they arguably don’t meet one of CN’s criteria on technical grounds.

AMF pretty clearly meets all of CN’s published criteria except:

Spending Practices: We exclude charities that report $0 in fundraising expenses, as we are interested only in charities that actively solicit donations from the general public. If you look at AMF’s most recent Form 990, they report zero fundraising expenses. On the other hand it’s clear that they actively solicit donations from the general public through their website. Indeed, thanks to their commendable transparency, you can see they get dozens of donations a day. So I’d argue that AMF meets the spirit but not the letter of this requirement. On their website, AMF states that as of October 2012, CN was citing this as a reason not to review them. • Scott on April 27, 2015 at 3:34 pm said: Any thoughts about helping develop space elevator technology? It would likely be a critical step in preventing many total annihilation scenarios by helping settle other planets. • What is your primary constraint to growth? (Among your traditional line of work). My guess is that you would identify your constraint as room for more funding among recommended charities. That is, you think you could relatively easily mobilize more dollars if you identified more high quality room for funding. Your impact is thus defined as money moved into high quality organizations with room for funding. Reading your disaster relief post, it seems that there is another avenue for impact: moving money out of low quality options and into medium quality or high quality with limited room for more funding. Would you ever consider increasing your marketing efforts in an effort to grow the market for effective altruism? In this scenario, you would achieve impact by moving money into marginally better opportunities. • Aceso Under Glass on April 28, 2015 at 7:12 pm said: https://docs.google.com/spreadsheets/d/1NpcmVO3DnOmtKRbYGkNbR0NTitclT5mwGBwBEwYlKHg/edit#gid=0 refers to low-skill occupational licensing, but the linked documents refer to high-skill occupations like medicine. The cost/benefit analysis seems it would be pretty different between the two. Is there other research into low skill occupations you didn’t publish? • Colin Rust on May 2, 2015 at 11:10 am said: A technical comment re Holden’s discussion of Cluster vs. Sequence thinking (a post I’m a big fan of): I’m not sure taking the geometric mean of the probability densities is an appropriate approach. It’s not associative in the following sense: If you have 3 distributions A,B,C and you combine A with B, then combine the combination with C, you get a different result than if you combine B with C and then combine with A. So for example if you have three independent normal distributions: A=N(0,1), B=N(0,1) and C=N(1,1) then you would surely expect the combined EV to be 1/3 (and indeed that’s what you get with any order under the normalized product methodology). But with the geometric mean methodology first combining A with B gives an EV of 1/2, whereas first combining B and C gives 1/4. • Neville Jones on May 4, 2015 at 8:55 am said: I have long believed in the need for a charity fund rating agency that would operate in a similar way that other ratings agencies (like Standard & Poors) rate the financial performance of other funds. So Bravo to GiveWell! I became aware of GiveWell this evening when highly regarded ethicist Peter Singer effectively endorsed GiveWell on national TV program Q&A. So Bravo again! Following the endorsement chain, I am thrilled that AMF is on your very short list of your top charities. I travel to Africa regularly and I am only too aware of the misery that Malaria brings to that continent. So here is my question. I was somewhat dismayed by your analysis of the “cost per life saved” analysis. I did a similar very rough calculation a few years ago and came up with a very different number. Less than$1,000 in fact.

And then I read that you completely discounted lives saved in the 5-14 year cohort. That is, to summarise simply, you divided the cost of distributing nets to people of all ages, by the number of lives saved in the under-5 cohort to arrive at the cost per life saved. Why?

You do go on to say that there may be some additional benefits etc but I dont understand why lives saved the 5-14 group was ignored. Especially since all the data is in your spreadsheet, except notably the “Deaths averted per protected child” in the 5-14 cohort. So, if we assume that the death aversion rate is the same for 5-14’s as it is for under 5’s (why wouldn’t it be?) the “cost per life saved” is then $891. Somewhat closer to my thumbnail estimate. Why is it not valid to count lives saved in the 5-14 cohort? Neville • 27chaos on May 6, 2015 at 4:42 pm said: Have you ever considered in detail devoting effort to helping Native American reservations? I did a quick search of earlier posts, but all I found was this: https://blog.givewell.org/2008/07/01/a-unique-giving-opportunity/ And I’m not sure if you even followed up on that? It occurred to me recently that the reservations are almost optimal for charity work because they’re much more accessible than other countries yet often still have a comparable quality of life. • Alexander on May 6, 2015 at 10:01 pm said: Thanks all for the comments and questions! Colin – on the bednet cost issue, I agree with your intuitive reading of that page and changed it to say$5. And I’m going to have to leave the technical comment to Holden to address 🙂

Scott – we haven’t investigated space elevator technology.

Ryan – we see our primary constraint to growth as management and, secondarily, staff capacity. We’re less interested in moving funds from bad to mediocre uses than driving funds to truly outstanding opportunities; we’ve written more about that decision here. We’ve considered investing more in marketing to grow our reach, and it’s not a top priority for us right now; other individuals and organizations are doing some work to help us in that vein.

Aceso Under Glass – we’re planning to write more about this soon, but we tentatively believe that the magnitude of the problems caused by higher-skill occupational licensing is significantly larger than those caused by lower-skill occupational licensing, though the latter may be more feasible to address.

Neville – The main reason we don’t include 5-14 year olds in the cost-effectiveness analysis is that the five randomized control trials estimating the mortality impacts of bednets focus only measure impacts in under-5 year olds (where the mortality impact is believed to be larger). We don’t know of any randomized control trials assessing the impact of bednets in older children, but we’d expect them to be smaller because there are fewer malaria deaths amongst older kids (and because researchers haven’t conducted RCTs that reported on malaria death impacts for them).

• Alexander on May 11, 2015 at 4:01 pm said:

27chaos – thanks for asking – we haven’t followed up on that.

• Neville Jones on May 11, 2015 at 9:22 pm said:

@Alexander – Thank you for the response. In that case I think it is misleading to include the cost of net provision to the 5-14 cohort in calculating the cost-per-life-saved. If the benefit is not being tallied in the cost-benefit analysis, then neither should the cost.

N.

• Alexander on May 12, 2015 at 1:38 pm said:

Neville – thanks for the thoughtful reply. I agree that we’re likely missing some of the benefits, but ignoring the costs of distributing the nets to most of the population (they’re distributed to everyone, not just those under 5) would more radically distort the estimated cost-effectiveness. To get an upper bound on how much our method of assuming no mortality impact on those over 5 would affect the cost-effectiveness calculation, we can look at what portion of all malaria deaths are attributable to kids under 5. The World Health Organization estimates that 77% of malaria deaths occur in children under 5 years old; another estimate implies that 38% of malaria deaths in Africa are amongst under-5 years olds. Taking the 38% estimate and aggressively (because there is no RCT evidence) assuming that bednets would prevent malaria deaths in adults at the same rate as for younger people would only yield an estimate for cost per life saved 1/.38 = ~2.5 times lower than our current estimate. Using the WHO figure instead of the other, more aggressive estimate and applying the same assumption would lead to a cost per life saved figure of 1/.77 = ~1.3 times lower than our current estimate. However, if we had to guess, we’d guess that our current estimates, which assume no impact of bednets on those over age 5, are more accurate than these ones that assume a proportionate impact, though of course we can’t be confident of that. In general, we’re not confident that our cost-effectiveness numbers aren’t off from reality by ~2x; there’s a lot of guesswork and assumptions involved.

• 27chaos on May 19, 2015 at 1:18 pm said:

Alexander – thanks for the reply. Will you please express to others at Givewell my opinion that Native American reservations seem like a potentially important charitable cause you have largely overlooked thus far?

Third world quality of living, with first world quality remedies potentially available. Easier to generate funds for, since people are biased to help causes close to home. And there’s even a sense of historical injustice and guilt that can be appealed to.

“We’re making things right again in this nation, for all its inhabitants. Nobody who lives in the United States of America should suffer like they’re living in the middle of a warzone.”

• 27chaos on May 19, 2015 at 1:18 pm said:

Also, it would be a boost to the US economy, probably.

• Deborah Adams on May 20, 2015 at 6:36 pm said:

What are the usual challenges that keep GiveWell from hiring a larger percentage of applicants (ex. lack of writing skills, quantitative skills, motivation)? Also what about for trial hires who aren’t given full-time offers?

• Ryan on May 20, 2015 at 6:46 pm said:

Thanks for the response, Alexander. I would be interested in more blog coverage of Deborah’s question, given that you have identified staff/management capacity as your primary constraint to impact.

• Alexander on May 20, 2015 at 9:20 pm said:

Thanks for the questions and comments!

27chaos – my impression is that Native American reservations aren’t likely to be good targets for the priority interventions our top charities carry out, but I’ll flag your comment for others on staff in case they disagree.

Deborah – we don’t track those statistics across our hiring processes in a uniform way, so I don’t have a quantitative answer. In general, we have high standards and the default is that we won’t hire someone who enters the process, rather than the converse. We try to assess candidates as a full package, so it’s hard to point to any one thing as a “deal breaker” in most cases.

Ryan – thanks for the suggestion. I’m not sure we have that much of interest to say on the topic, but we’ll keep it in mind.

• Deborah Adams on May 21, 2015 at 5:53 pm said:

Thank you Alexander. I do agree with Ryan’s consideration. If hiring is a major difficulty in building GiveWell’s impact, then I would think it is worth giving rigorous thought to. Perhaps GiveWell is rejecting some people when a more thoughtful assessment would indicate they could still be valuable assets or even change the areas that GiveWell dislikes (ex. writing style).

• Nicholas on June 24, 2015 at 1:10 am said:

Is there an effective charity that subsidizes highly effective contraception for the global poor?

I don’t mean “sex education” charities or anything like that. Just make contraception cheaper for those who have a demand for it?

• Ian Turner on June 24, 2015 at 6:16 am said:

Nicholas: Givewell has recommended Population Services International (PSI) in the past, a large charity which focuses mainly on contraception, AIDS prevention, and malaria.

https://www.givewell.org/international/charities/PSI