Afaict the model doesn’t apply well to effective altruists as a sub-population.
Most EAs are young, and unless the movement suddenly becomes popular among older people, few are likely to die in the next 40 years[1]. Even if you fully buy this type of argument, and you believe donation opportunities are roughly uniformly distributed over time, the members of the community should probably do some ‘donation smoothing’ while alive.
Ok so is that the only weakness you see in the model? If so, what if we look at the entire population of donators, rather than the “EA population”. Or just do the “smoothing”? My question just has to do with pure effectiveness
I don’t think it’s specifically about the EA population.
The value of donations may change over time. Your model shows that investing results in having more money to donate in the future. But it doesn’t seem to take into account the value of that money (or the value of what the money can buy). This might change over time.
A couple of examples:
Assume everyone invests now rather than donates.
But in the year 2035 humanity is destroyed .
The value of the investments then becomes zero. We would have been better off donating to help prevent whatever caused humanity’s destruction.
(Tobias makes this point in his footnote above - ‘If there’s an existential catastrophe, donating at death won’t matter anyway’.)
Assume the world continues to get richer, and the poorest benefit the most. The world’s poorest increase their income from £1000 per year now to £4000 in 20 years time.
The value of your investments double in this time, so you are now able to donate £2000 to someone instead of £1000.
But donating £2000 in the future to someone who has £4000 is plausibly less valuable than donating £1000 today to someone who only has £1000.
This also relates to ogtutzauer and Henry Howard’s points—the sooner the recipient receives a donation, the faster they can invest it.
Those are cases where the value of the donation declines. There could be other cases where the value of donation increases—perhaps as a result of new research showing how resources can be used more effectively, or a new treatment etc. In this case, there would be more money in the future (due to returns on investment) and more cost-effective things to spend this money on.
I think we’re people seem to be getting tripped up with this is that everybody is not doing one thing or another in the sense that if everybody is giving at death, there is a steady stream of people dying so there is giving happening all the time. Does that make sense?
No, I don’t understand your point. The fact that some people are dying and donating now doesn’t answer the question of whether people who are not dying now should donate or invest now.
If there is a population of let’s say 10 million people that are donators, then there is a certain amount of money per unit time that is being donated. To the recipients of the donation, it would matter less the individual behavior is than the behavior of the population is. So if there is a model whereby holding onto one’s wealth (and being motivated to grow it and enjoy it) results in more money per unit time donated within the population, that would seem to be more effective to me. So , assuming some sort of coordination or at least consensus about the best way to give, is based on not just that one person
I think we may be going round and round. The amount is related to the value by some constant, unless you’re implying a time factor, but as far as I can see, the timing concern is a red herring, as in my model the pump is already primed and wealthy enough donors are dying in week one.
The amount is related to the value by some constant, unless you’re implying a time factor
Yes, I and others have explicitly been saying time is a factor, e.g. see my 2 examples above and comments by others. The amount is likely not related to value by some constant. The value that a given amount of money can buy will vary over time.
but as far as I can see, the timing concern is a red herring, as in my model the pump is already primed and wealthy enough donors are dying in week one.
I’ll be honest, I don’t understand this point, or why it means the value of a donation won’t change over time.
I may just be misunderstanding. But I don’t think there’s much more to say on this, unless e.g. you’re able to share your model on Excel or Google sheets.
One thing is that ‘donations at death’ are likely qualitatively different to ‘donations during life’. This difference might be important in terms of how effective/desirable the individual donations are.
Currently, I’d expect older people to donate less to animal welfare than younger people. In the future, the way donation patterns differ might change. But there might be some differences that stay consistent over time, e.g. older people donating in a more/less risk-averse manner.
I didn’t take that into account! So yeah we have to factor in an assumption that the more we advance the average age of the giver, the less progressive the causes, with a bit of “lag” expected, for better or worse!
Afaict the model doesn’t apply well to effective altruists as a sub-population.
Most EAs are young, and unless the movement suddenly becomes popular among older people, few are likely to die in the next 40 years[1]. Even if you fully buy this type of argument, and you believe donation opportunities are roughly uniformly distributed over time, the members of the community should probably do some ‘donation smoothing’ while alive.
(If there’s an existential catastrophe, donating at death won’t matter anyway.)
Ok so is that the only weakness you see in the model? If so, what if we look at the entire population of donators, rather than the “EA population”. Or just do the “smoothing”? My question just has to do with pure effectiveness
I don’t think it’s specifically about the EA population.
The value of donations may change over time. Your model shows that investing results in having more money to donate in the future. But it doesn’t seem to take into account the value of that money (or the value of what the money can buy). This might change over time.
A couple of examples:
Assume everyone invests now rather than donates.
But in the year 2035 humanity is destroyed .
The value of the investments then becomes zero. We would have been better off donating to help prevent whatever caused humanity’s destruction.
(Tobias makes this point in his footnote above - ‘If there’s an existential catastrophe, donating at death won’t matter anyway’.)
Assume the world continues to get richer, and the poorest benefit the most. The world’s poorest increase their income from £1000 per year now to £4000 in 20 years time.
The value of your investments double in this time, so you are now able to donate £2000 to someone instead of £1000.
But donating £2000 in the future to someone who has £4000 is plausibly less valuable than donating £1000 today to someone who only has £1000.
This also relates to ogtutzauer and Henry Howard’s points—the sooner the recipient receives a donation, the faster they can invest it.
Those are cases where the value of the donation declines. There could be other cases where the value of donation increases—perhaps as a result of new research showing how resources can be used more effectively, or a new treatment etc. In this case, there would be more money in the future (due to returns on investment) and more cost-effective things to spend this money on.
I think we’re people seem to be getting tripped up with this is that everybody is not doing one thing or another in the sense that if everybody is giving at death, there is a steady stream of people dying so there is giving happening all the time. Does that make sense?
No, I don’t understand your point. The fact that some people are dying and donating now doesn’t answer the question of whether people who are not dying now should donate or invest now.
If there is a population of let’s say 10 million people that are donators, then there is a certain amount of money per unit time that is being donated. To the recipients of the donation, it would matter less the individual behavior is than the behavior of the population is. So if there is a model whereby holding onto one’s wealth (and being motivated to grow it and enjoy it) results in more money per unit time donated within the population, that would seem to be more effective to me. So , assuming some sort of coordination or at least consensus about the best way to give, is based on not just that one person
This is still only considering the amount of money donated, not the value of what the money can buy.
I think we may be going round and round. The amount is related to the value by some constant, unless you’re implying a time factor, but as far as I can see, the timing concern is a red herring, as in my model the pump is already primed and wealthy enough donors are dying in week one.
Yes, I and others have explicitly been saying time is a factor, e.g. see my 2 examples above and comments by others. The amount is likely not related to value by some constant. The value that a given amount of money can buy will vary over time.
I’ll be honest, I don’t understand this point, or why it means the value of a donation won’t change over time.
I may just be misunderstanding. But I don’t think there’s much more to say on this, unless e.g. you’re able to share your model on Excel or Google sheets.
(I haven’t looked deeply into the model.)
One thing is that ‘donations at death’ are likely qualitatively different to ‘donations during life’. This difference might be important in terms of how effective/desirable the individual donations are.
Can you elaborate a little bit? I wanna see what your saying.
Currently, I’d expect older people to donate less to animal welfare than younger people. In the future, the way donation patterns differ might change. But there might be some differences that stay consistent over time, e.g. older people donating in a more/less risk-averse manner.
I didn’t take that into account! So yeah we have to factor in an assumption that the more we advance the average age of the giver, the less progressive the causes, with a bit of “lag” expected, for better or worse!