Providing insurance is really hard, and insurance coverage is inherently limited. From the LessWrong article I linked:
This sounds a lot like insurance, and effective altruists could indeed insure themselves against a myriad of risks. Unfortunately, insurers will reject claims whenever possible, many risks cannot be easily insured, and insurance requires monthly payment, which requires a stable income. In other words, insurance is no replacement for a personal financial buffer. Insurance is useful to mitigate specific high-cost risks, but that’s it.
(Additionally, insurance introduces the overhead of an insurance company. When I asked for a ballpark estimate at the insurer I interned at, I was told that at most one-third of all premium money is put back into payment of legitimate claims.)
Regarding the specifics, I’m not really clear on why an emergency fund should be coupled with donations.
By focusing exclusively on reimbursing donors in financial trouble, we avoid opening a can of worms. First of all, the risk of fraud is much lower. If you can only get back half of what you gave away, there is no way to use the fund to make money, unless you control a GiveWell-recommended charity. Second, we do not have to judge whether someone is EA-aligned. Third, people cannot take advantage of the fund and, perhaps more importantly, people will not have to worry other people are taking advantage of the fund.
To expand on the last point, if we ever decide to reimburse a donation because someone’s second car broke down, that might annoy some people with a different idea of what constitutes an emergency, but at least they know that person donated at least twice the amount needed to fix the car. Now, if we handed out fix-your-second-car money to people who never donated anything to charity, I’d predict riots.
Also, coupling the two promotes donations. People who would normally find parting with substantial amounts of money scary have the assurance that they can always knock on our door.
1. Cap the size of the reimbursement fund at some smaller portion than 10% of total receipts over time, and
2. Request that recipients repay their reimbursement, perhaps suggesting that instead of donating 10% of income going forward, they give through the new org, and put 50% of donations to reimburse their previously refunded payment and 50% to charity.
3. Disclose the payments to individuals with a 1-year lag unless the payment to them has been repaid, with the option for recipients to request a longer lag to announce for them to repay if they don’t want it announced.
As Linch says, this is a vision for this new fund that is different than what the OP and what I think sentiment supports.
In one generous vision, we have a institution where someone can donate most of their income each year. They can trust that if their loved one gets sick, they can pull out a large fraction of the money to help, like 99.99% of humanity. It’s none of our business if they do this. It’s a feature.
There is an aesthetic among older EAs (not carpet bagging arrivistes like myself) of donating huge fractions of their income, kidneys, etc. Some of these people are in great positions—some aren’t and I’ve heard of anecdotes where early EAs can get burnt out and in difficult situations later.
If you believe the above, a 10% limit and a public announcement of both their “clawback” (and implicitly their personal situation) is really different. I would guess this thins the value of this project to the degree that it’s not worth the legal and logistical challenges which are pretty big (there’s even a mechanism design sort of challenge).
I’d think that saying who was given money would be a pretty fundamental part of transparency for this, as is the expectation that people would continue giving once whatever emergency exists is resolved. Yes, 1 year is likely too short a time frame to repay if there is a huge need, but that’s why I’d say people should be able to suggest a longer time frame. (And if someone has a huge disaster and needs cash, I’d think that they wouldn’t mind people hearing that fact quite as much.)
The use case, in my mind, is someone who is nervous about giving 25% of their income and not trying to keep 6 months of savings in case of disaster. If they did that, then something happens, they need to change their mind about some of the giving—but once they are back on their feet, they would go back to donating. However, if it’s building a way for people to later say “I want half my money back, to keep forever, now that I no longer want to do this,” I don’t think it’s worthwhile as a project.
That seems plausible—but I don’t know how this would work if there wasn’t any transparency about who was given money back, so I’m not sure how to avoid #3.
And For #2, I think the point is that this isn’t a payout for disaster, it’s a temporary solution to replace savings. Once they are back on their feet and able to donate, I’d think that paying a larger part into the fund makes sense. But I agree that it’s a different vision than using this as pure insurance.
Providing insurance is really hard, and insurance coverage is inherently limited. From the LessWrong article I linked:
By focusing exclusively on reimbursing donors in financial trouble, we avoid opening a can of worms. First of all, the risk of fraud is much lower. If you can only get back half of what you gave away, there is no way to use the fund to make money, unless you control a GiveWell-recommended charity. Second, we do not have to judge whether someone is EA-aligned. Third, people cannot take advantage of the fund and, perhaps more importantly, people will not have to worry other people are taking advantage of the fund.
To expand on the last point, if we ever decide to reimburse a donation because someone’s second car broke down, that might annoy some people with a different idea of what constitutes an emergency, but at least they know that person donated at least twice the amount needed to fix the car. Now, if we handed out fix-your-second-car money to people who never donated anything to charity, I’d predict riots.
Also, coupling the two promotes donations. People who would normally find parting with substantial amounts of money scary have the assurance that they can always knock on our door.
I’m normally skeptical of “emergency fund” ideas in large part because it’s hard to decide “who counts,” so I like that this solves that problem.
I think you should probably also:
1. Cap the size of the reimbursement fund at some smaller portion than 10% of total receipts over time, and
2. Request that recipients repay their reimbursement, perhaps suggesting that instead of donating 10% of income going forward, they give through the new org, and put 50% of donations to reimburse their previously refunded payment and 50% to charity.
3. Disclose the payments to individuals with a 1-year lag unless the payment to them has been repaid, with the option for recipients to request a longer lag to announce for them to repay if they don’t want it announced.
As Linch says, this is a vision for this new fund that is different than what the OP and what I think sentiment supports.
In one generous vision, we have a institution where someone can donate most of their income each year. They can trust that if their loved one gets sick, they can pull out a large fraction of the money to help, like 99.99% of humanity. It’s none of our business if they do this. It’s a feature.
There is an aesthetic among older EAs (not carpet bagging arrivistes like myself) of donating huge fractions of their income, kidneys, etc. Some of these people are in great positions—some aren’t and I’ve heard of anecdotes where early EAs can get burnt out and in difficult situations later.
If you believe the above, a 10% limit and a public announcement of both their “clawback” (and implicitly their personal situation) is really different. I would guess this thins the value of this project to the degree that it’s not worth the legal and logistical challenges which are pretty big (there’s even a mechanism design sort of challenge).
I’d think that saying who was given money would be a pretty fundamental part of transparency for this, as is the expectation that people would continue giving once whatever emergency exists is resolved. Yes, 1 year is likely too short a time frame to repay if there is a huge need, but that’s why I’d say people should be able to suggest a longer time frame. (And if someone has a huge disaster and needs cash, I’d think that they wouldn’t mind people hearing that fact quite as much.)
The use case, in my mind, is someone who is nervous about giving 25% of their income and not trying to keep 6 months of savings in case of disaster. If they did that, then something happens, they need to change their mind about some of the giving—but once they are back on their feet, they would go back to donating. However, if it’s building a way for people to later say “I want half my money back, to keep forever, now that I no longer want to do this,” I don’t think it’s worthwhile as a project.
A lot of these choices seem unnecessarily punitive to me, not sure.
That seems plausible—but I don’t know how this would work if there wasn’t any transparency about who was given money back, so I’m not sure how to avoid #3.
And For #2, I think the point is that this isn’t a payout for disaster, it’s a temporary solution to replace savings. Once they are back on their feet and able to donate, I’d think that paying a larger part into the fund makes sense. But I agree that it’s a different vision than using this as pure insurance.