And because, I suspect, the actual cost-effectiveness is often likely very low, even a highly uncertain CEA would point to that, and thus excuses like âitâs too complexâ are used to avoid having to confront that.
I am instantly more trustful of any charity or cause area that attempts to estimate their cost-effectiveness, especially if they have strong reasoning transparency and highlight their key uncertainties and suspected error bars.
AIM /â Charity Entrepreneurship, for example, estimates cost-effectiveness for advocacy ideasâthese are extremely hard to do accurately and have massive uncertainties. They are also very much reliant on others and it is impossible to specifically isolate the charityâs impactânonetheless, they provide a transparent, good-faith estimate. (e.g.)
I am normally a bit distrustful of any cause areas or charities that claim it is too difficult to quantify impact or calculate cost-effectiveness.
Phrases like âthe work is complexâ or âdata is difficult to collectâ are kinda red flags to me, not because they are untrue, but because:
they are also true of a huge number of charities and cause areas for which EAs are able to make estimates. (These estimates might be super uncertain, and have big error bars, but they are better than nothing).
And because, I suspect, the actual cost-effectiveness is often likely very low, even a highly uncertain CEA would point to that, and thus excuses like âitâs too complexâ are used to avoid having to confront that.
I am instantly more trustful of any charity or cause area that attempts to estimate their cost-effectiveness, especially if they have strong reasoning transparency and highlight their key uncertainties and suspected error bars.
AIM /â Charity Entrepreneurship, for example, estimates cost-effectiveness for advocacy ideasâthese are extremely hard to do accurately and have massive uncertainties. They are also very much reliant on others and it is impossible to specifically isolate the charityâs impactânonetheless, they provide a transparent, good-faith estimate. (e.g.)