It’s possible that, as Phil suggests, talking about bad charities will make people realize the importance of being selective about where you give. But ultimately seeking out bad charities is a pointless exercise: The vast majority of charities (of anything, really) are mediocre, so if you avoid bad charities then you are still likely to end up giving to a mediocre organization. Only by seeking out the spectacular can you obtain stunning results.

If you don’t discuss bad charities, people may not realize the importance of seeking out good charities.

Let’s assume that we could rank-order charities based on social utility created per dollar donated.

If we identify one set of donors who like Holden, are somewhat pessimistic about charities on the whole (i.e. many of them are ineffective are worse), then the spread in social utility between an elite organization (at the 90th percentile or better) and a very poor one (at the 10th percentile or worse) is quite wide. There is considerable value in an impact-oriented donor really studying things.

On the other hand, there may be a hypothetical optimistic donor, who thinks that most charities are well intentioned, and able to translate those good intentions into useful social impact, such that the spread between a top charity and a much weaker charity is not that great. To this donor, expending a lot of time researching which charities to donate to is not particularly important.

i.e. Spending at least some time discussing charities that appear to be ineffective helps understand the spread from a strong charity to a weak one, and thus, the importance of avoiding weak charities and hopefully finding strong charities.

The point of this post was not that we see no value in discussing the bad (we are gradually realizing how much value there is), but that for us personally the good is more interesting/important and this fact points to an interesting difference in priors.

There seems to be value in reporting on both the best AND the worst, especially different folk have different Bayesian priors. Some will be encouraged to donate to “the best”. Others may wish to mostly refer to their own counsel, but at least they may be convinced to avoid “the worst”.

Even in the unpleasant scenario that this just means that charitable dollars are going to a charity that is doing less harm than “the worst”, you’re still getting a better return on giving than if the info on “the worst” hadn’t been around.

So isn’t there some value in a “stinkers” list?

• Funders of smaller/still-developing charities often have a lot of information about them - different information from what we generally use, but valuable (arguably more valuable) information that comes from personal interactions with founders, firsthand observation, etc.
• Turning a smaller/developing charity into a proven, high-impact, scaled-up charity is a substantial upside that can justify a substantial risk. Investing in a small charity can therefore be reasonable even if the odds of failure are high - though we don’t recommend such an investment for individual/casual donors.

Those, to me, are the reasons that investing in smaller/developing charities makes sense. However, they don’t relate to the question of what your “prior” should be for a charity about which you have no information. (As a side point, I agree with James that “large and unproven” implies different things from “small and unproven,” though I’d put the “default assumption” as a low probability of success for either.)

Brian, one thing I don’t think you’re accounting for is the opportunity cost of human resources imposed by giving. We’ll discuss in a future post. It’s also good to keep in mind that many of the “gifts” provided by charities are in forms such that they cannot be sold or transferred (and when they can be sold or transferred, that additional concerns enter the picture).

James Edward Dillard and Ian Turner, it’s an interesting point and a tough call regarding foundations. Our thinking to date has been that we’re better positioned to change incentives for charities. Charities are asking for your money and foundations aren’t, and so encouraging people to give to only the most transparent charities creates a very clear link between rewards and good behavior. We have very little sense of what foundation incentives are. And arguably foundations do not have the same responsibility to share information, since they’re not soliciting.

That said, there are good reasons to argue that the current level of opacity at foundations is unacceptable. And it’s possible that they would be quite sensitive to public criticism, since their public image is one of the few concrete ways many seem to measure success. It’s something to think about.

Bill Biggs, charities doing more measurable things are more likely to be able to learn from their mistakes, more likely to be held accountable by donors, and thus - in our view - more likely to succeed. If the goal is to maximize good accomplished rather than to be “fair” to charities on their own terms, then we don’t think it makes sense to “sector-adjust” how strong the case for a charity is. That said, some sectors with less measurable short-term impact could have higher upside, and thus might be worth more risk. We’ll be discussing this point in the future.

Doug S, we prefer not to comment on charged political issues like this one when there isn’t a clear benefit to doing so, but assume for the sake of argument that Prof. Hanson is correct. There is still a good reason to focus on charity, which is that any individual can apply rationality to their giving and accomplish more good. By contrast, agreeing with Prof. Hanson on immigration would still leave you a long way from doing anything to improve the situation. This argument would not apply to a specific person with significant political power, but it applies to the layperson. Some semi-related thoughts at this very old post on my attitude toward politics.

Yi-An, your comment seems to imply that you agree that “no data” = “bad, very bad” (at least as a candidate for an individual/casual donor’s gift). I think that giving 0/3 stars is the best way to accurately communicate this. Our ratings are always accompanied by a “What do these ratings mean?” link.

I think perhaps this is a framing issue - you give nonprofits with not enough data 0 out of 3 stars, which implies they’re doing a terrible job (failing). A more accurate rating would be a separate category for “not enough data” - which would more truthfully reflect your own opinion that not all charities with no data are failing.

0 out of 3 stars should be reserved for nonprofits that are proven to have no impact or do harm (are failing). I recognize that such data is likely to go unpublished (or more likely, ungathered), but it’s more truthful. As to the “not enough data” category, it is a nuance - and you’re depending on your audience to be sophisticated enough to understand what that means (it’s bad, very bad). But if GiveWell is serious about transforming the culture of how nonprofits are viewed, then seek to live in the nuance and don’t blend “no impact” with “no data”.

I doubt it’ll stop people criticizing you, but hey, one step at a time.

Economist Robin Hanson, in an aside comment in a blog post on microlending:

Yet we’ve known for many decades how to help the world’s poor while actually benefiting us in the process: allow more immigration. Yes even immigration has limits, but that’s no excuse for why we haven’t done all we can.

Completely agree that Foundation nonsense and donor apathy are related and respect GiveWell’s focus.

My thought is that being “rougher” on Foundations might be a means to an end. As institutions, they are easier to find information on and have an interest in protecting their name (so will be more responsive to criticisms). Furthermore, because of their size, foundations have the ability to demand impact measurements, where they are feasible (thus making them available for all donors).

Thus by conducting an audit of 15-25 foundations that share one industry (say education) as a focus area, my thought is that GiveWell would be able to unearth a ton of information for impact minded donors (perhaps even ones who might want to contribute to a well-run foundation) in addition to a number of other benefits.

Does that make sense?

James

I donate to Transparency International. Very difficult to access their effectiveness simply by the nature of their work. A bad charity to me is one that is misleading, not transparent, or ineffective in comparison to its peers.