Yeah, that’s a good question. It’s underspecified, and depends on what your baseline is.
We might say “for $1 donated, how much can we increase consumption”. Or “for $1 donated, how much utility do we create?” The point isn’t really that it’s 10x or 5x, just that one opportunity is roughly 2x better than the other.
So if we are giving to, e.g., encourage policies that increase incomes for average Americans, we need to increase them by $100 for every $1 we spend to get as much benefit as just giving that $1 directly to GiveDirectly recipients.
That’s not exactly “Return on Investment”, but it’s a convenient shorthand.
So it’s like a benefit to cost ratio. So I can see with diminishing returns to more money, the benefit to cost ratio could be half. So with $1 million in the early days of EA, we could have $10 million of impact. But now that we have $1 billion, we can have $5 billion of impact. It seems like the latter scenario is still much better. Am I missing something?
No, I meant an intervention that could produce 10x ROI on $1M looked better than an intervention that could produce 5x ROI on $1B, and now the opposite is true (or should be).
Yeah, that’s a good question. It’s underspecified, and depends on what your baseline is.
We might say “for $1 donated, how much can we increase consumption”. Or “for $1 donated, how much utility do we create?” The point isn’t really that it’s 10x or 5x, just that one opportunity is roughly 2x better than the other.
https://www.openphilanthropy.org/blog/givewells-top-charities-are-increasingly-hard-beat
That’s not exactly “Return on Investment”, but it’s a convenient shorthand.
So it’s like a benefit to cost ratio. So I can see with diminishing returns to more money, the benefit to cost ratio could be half. So with $1 million in the early days of EA, we could have $10 million of impact. But now that we have $1 billion, we can have $5 billion of impact. It seems like the latter scenario is still much better. Am I missing something?
Uhh, I’m not sure if I’m misunderstanding or you are. My original point in the post was supposed to be that the current scenario is indeed better.
Ok, so we agree that having $1 billion is better despite diminishing returns. So I still don’t understand this statement:
Are you saying that in 2011, we would have preferred $1M over $1B? Or does “look better” just refer to the benefit to cost ratio?
I think I see the confusion.
No, I meant an intervention that could produce 10x ROI on $1M looked better than an intervention that could produce 5x ROI on $1B, and now the opposite is true (or should be).