[andy_ppp]

This is really interesting and a great idea; however, I'd almost trust charities more if they listed how they had failed each year and where delivery and services went wrong and what mitigation has been put in place.

It would be good to have a system where charities were allowed to fail like bad companies as well, but it's difficult to be this honest.

[tmacaulay]

Thanks, we completely agree!

Right now, the incentives in the charity sector are totally screwed. The charities which get the biggest are the ones with the best marketing, not the ones which have the most impact. That means that big charities often hide their mistakes and focus on maintaining a wholesome image.

In contrast, many of the organizations we work with regularly publish their mistakes and lessons. By sharing this kind of information, the whole sector can learn and improve. In particular, GiveWell (http://www.givewell.org/about/our-mistakes) has a great page listing their mistakes.

We've been especially impressed with charities like New Incentives, who realised that the original target population they were trying to serve (pregnant women with HIV) wasn't big enough, so they pivoted to focus on incentivizing mothers to vaccinate their children so that they could gather more evidence and have an even bigger impact.

With EA Funds, donors pick which problems they want their donation to solve, and we find the best giving opportunities. We will fund both new start-up style charities and larger more validated approaches. We will fund charities which have run failed programs in the past, provided that they have updated their approach.

[ArneBab]

So New Incentives changed what they wanted to do because they would get paid more that way?

I know this sounds harsh. This is the incentive you create: Pushing people who initially wanted to help where they see need to instead focus on helping where they get more money for do so.

And a sufficient amount of free money is mostly available to a very small group of people.

This does not mean it’s necessarily bad. It just means that its incentives are skewed, too: The charities are pushed to become interest groups of the rich (to some degree this is also the case today, but stronger quality assurance also means more control to follow the largest donors' wishes).

[willmacaskill]

I think that's exactly wrong, I'm afraid.

In traditional charity, the incentives that charities have are to do whatever is going to fundraise the most. So the charities that get biggest are those that are best at looking good, rather than doing good.

The solution is to have a set of donors who really care about funding whatever does the most good. That means that a charity's fundraising incentives line up with what's actually best for the world. And that set of impact-motivated donors is exactly what we're trying to create with the effective altruism movement.

New Incentives is a great example of that working. In just the same way that a startup will pivot if it thinks it could be working on something else that's more profitable, because of the existence of the EA community New Incentives is able to pivot to a different approach that it thinks will do more good per dollar and knows that, if it succeeds at doing that, it will be able to grow more.

[tmacaulay]

Thanks for your feedback, the dynamic you're talking about is exactly the problem we're trying to solve. We want the best charities, which target the biggest needs, in a cost-effective way, to get the most funding.

New Incentives changed what they were focusing on because their initial program was proven to be less effective. They got some initial results from their study, and it showed that because they could only reach a smaller number of people, their program didn't hit the threshold they were targeting for cost-effectiveness.

Right now charities are incentivised to skew their programs towards areas in which they can get the most funding. We're trying to fund programs based on effectiveness, and build a community of donors who will donate to whatever programs are proven to be the most effective. If people founding non-profits know that there is a community of donors who will fund programs that work, we fix these incentives, and we hope to see many more effective programs launched. This should make it easier for effective charities to get the funding they need to grow.

We don't focus on special interest groups, and we don't fund whatever our largest donors are most interested in, we only fund programs that are highly effective.

[Analemma_]

The charities themselves rarely do that directly (unfortunately, as the OP said, the incentives are really messed-up so they would never do so), but GiveWell does it "by proxy" every year, so if you donate to GiveWell's "general fund" you can be sure that they're iterating and correcting for failures.

In terms of other feedback, I have to say I really like the clean separation of "ends" and how you can pick which one(s) you want. TBH, I've been getting increasingly annoyed at how the X-risk people are gradually taking over EA, but I don't want to just quit or blow up at them either (they mean well even if I think they're misguided). This looks like a way that everyone can be happy. Please don't let them needle you into adding that stuff into the Global Development or Animal Welfare funds.

[bradleyjg]

I agree with you about X-risk, but I'm not sure the separation is so clean. For example looking at the EA Community fund page it isn't at all clear to me whether or not some of the money will end up in the hands of the AI Safety people. To be fair this is openly disclosed as a risk, which is great transparency.

Second, donors may choose not to support the Movement Building Fund if they do not wish to indirectly support all of the problem areas that the EA community is likely to support. This includes areas like Global Health and Development, Animal Welfare, Long-Term Future, and any future problem areas that the community deems effective to address. For example, Dylan Matthews criticized the EA community in 2015 for being overly self-promoting and overly concerned about risks from advanced artificial intelligence (one response to this criticism here). Those with strong views about which problem areas they do not wish to support might avoid movement building as a result.

However, admirable transparency or not, the only fund I'd really be interested in allocating money to would be the Global Health and Development Fund. Given that, for me personally, the paragraphs about differentiating vs GiveWell are critical.

[tmacaulay]

It's really great to hear this view. While lots of donors split their donation between our funds, many people choose to allocate 100% of their donation to a single fund. At a later date, we want to allow people to share their fund allocations, so donors can compare allocations and discuss differences in cause prioritization.

Right now, our Global Health and Development Fund has 61% of all donations by value, with the Long-Term future fund coming in at 22%. It will be exciting to see how this changes over time, and whether there are differences in fund allocation by geographic area or demographics.

[colophonemes]

For sure, we want to represent a broad range of views within areas that are potentially high impact. We chose these as our initial funds partly because they seem like cause areas that reasonable people can disagree on (both on questions of values, and on empirical questions about relative risks and how to solve problems).

[kyleschiller]

Yeah! Givewell does exactly that

http://www.givewell.org/about/our-mistakes