Anonymous asked: Do you agree with the general opinion shared by many people associated with GWWC that diversifying donations is bad and you should go for your best pick? My impression is that diversification is better due to the high uncertainty levels, plus it feels easier to adhere to pledges if your money supports a wide range of charities and you're more likely to produce a positive outcome.

Personally, I am pretty confused by the widespread sense in EA that there is no good rational argument for diversifying one’s individual charity donations.

Arguments against individual diversification tend to rely on ignoring effects on things other than the direct work done by charities, like effects on incentives for charities (to become more effective, or do better measurement/reporting/research).  On a truly individual level, maybe these effects really are negligible.  But these discussions aren’t about what one person should do, they’re about EA norms that – if the person making the argument had their way – would be adopted by many people.

So to make this kind of argument work, one would need to argue that the entirety of EA (or of EA except some tiny set of mega-donors) does not move enough money to have non-negligible indirect effects.  For all I know, this may have some truth to it, but I never see it being claimed in the course of anti-diversification arguments.  Instead there’s just a step left missing.

This is related to a larger, interesting issue about the role of EA: is it just about observing the charities that currently exist, selecting the best of the pool (no matter how suboptimal it may be relative to charities that could exist), and spamming money at it?  Or is it also about rewarding standards of measurement, scientific practice, self-improvement, etc. among charities?

If it’s the latter, that introduces various considerations about rewarding good practice even in the presence of not-quite-best-in-show outcomes, but these will be moot if (as anti-diversification arguments seem to assume) the entirety of EA can’t incentivize charity behavior to any appreciable extent.  If it actually can’t, that seems like a very important fact with major negative implications for some visions of EA, and a much bigger deal than individual diversification decisions.

FWIW, here’s a quote from GiveWell’s website that seems like a counter-example to the claim that aggregate EA donation patterns can’t affect charity behavior:

slatestarscratchpad:

I think it’s technically correct that you shouldn’t diversify. Uncertainty doesn’t matter unless you are personally providing a substantial amount of the world’s charity budget.

I agree that emotionally it’s much easier to diversify and in practice I diversify a lot.

We follow up intensively with our top charities over time, and consider this one of the major arguments in favor of supporting such charities. Because our recommendation directs substantial donations to a charity, top charities are generally willing to engage substantively with us and help us deepen our understanding of their activities and progress over time.