One thing Iâd like to add is that when thereâs pushback on donation matching initiatives, the discussion often focuses on âis the communication of what happens clear?â and âare donors misled?â and so on. I think (lack of) honesty and clarity is the biggest problem with most donor matches, but even where those problems are resolved, I think there are still other problems (e.g. it just seems kind of uncooperative /â adversarial to refuse to do something good that youâre willing and able to do unless someone else does what you want). My favoured outcome is not âhonest matchingâ but âno matchingâ, and Iâd like the discussion to at least have that on the table as a possible conclusion.
(To be clear, I think neither Giving Multiplier nor FarmKind are intending to be dishonest, although I think itâs still unclear to me whether they are unintentionally misleading people. I perhaps think that they are applying ordinary standards of truthfulness and care, whereas Iâd like us to have extraordinary standards.)
If providing funds that will contribute to a match has the effect of increasing funds generated to effective charities and is transparent and forthright about the process involved, I donât really see the problem.
It does not seem to me as an uncooperative and adversarial process, but one that gets more people involved in effective giving in a collaborative and fun experience. As the OP suggested, there is a concern that the matcher (as opposed to the fund provider) may be deceived and not realize that the net effect of their participation is less funds to their preferred charity and more to the fund providerâs preferred charity. But if there is adequate understanding of the process, it seems to me that the question boils down to whether or not the availability of Giving Multiplier or Farmkind results in more counterfactual donations going to these charities.
If providing funds that will contribute to a match has the effect of increasing funds generated to effective charities and is transparent and forthright about the process involved, I donât really see the problem.
I think this is a pretty interesting aspect of the discussion, and I can see why people would not only agree with this but think it kind of obvious. Here are some reasons why I donât think itâs so obvious:
I think itâs worth noticing when you offer other people something that you wouldnât take yourself. Why is it good for them, but not good for you?
I think you should include your assessment of what value the deal has to someone else in your assessment of whether you have successfully communicated what the deal is. As above, if people are taking an offer that you wouldnât take, and that as far as you see doesnât benefit them, I think you should take that as evidence that you havenât explained the deal (Jason makes this point in another comment)
I think a useful analogy is a casino, say a roulette wheel. The rules of roulette wheels are pretty simple, and completely public. People play them willingly, often believing itâs in their interests to do so. Yet I think roulette wheels are kind of exploitative and even verge on predatory, and I wouldnât feel comfortable running one even to donate the proceeds to charity. Again, opinions can differ here, but for me I see facilitating and encouraging other people to make decisions that you know are bad isnât ethical behaviour, even if they ultimately make the decision freely and willingly. (To be clear, I think offering a donation match is less unethical than running a casino, but I bring up the casino to illustrate how offering people a free and informed choice sometimes still doesnât sit right with me.)
I think your example of the roulette wheel is illustrative⊠People know that they are going to lose money in expectation but it is rational for many of them to play because the expected fun they have from a night at the casino exceeds the expected loss. For example if you churn $500 and the average payback is 92.5%, the expected loss is $33.50 and you might have more fun than with other recreational activities that cost that much. Of course, some people may have gambling problems or may irrationally think that they have an advantage. These are the cases that we should be worried about: people who have issues of understanding or volitional capacity.
Similarly, people can understand whatâs going on with Farmkind or Giving Multiplier and want to participate, understanding that their choice is benefiting the charities that Farmkind or Giving Multiplier prefer.
If they enjoy and understand the process employed, I donât understand why one would think that they were harmed or exploited
On the casino example: I agree that a gambler could decide that the expected financial loss was outweighed by the value they ascribed to the entertainment. However, that doesnât tell us what proportion of people around the roulette wheel fall into that category vs. people with impulse-control disorders or the like vs. people who do not have disorders yet make harmful decisions because they are too influenced by gamblerâs fallacies. I would put the burden on the would-be casino operator to prove that enough[1] of their would-be customers fell into the ârational gamblerâ category before considering Benâs objection to running a roulette wheel have been satisfactorily addressed.
Turning to the analogy, I think there is considerably more evidence of the ârational gamblerâ category existing than the âpeople who truly understand an FarmKind-type match, but enjoy it enough to experience it as a net positive.â A full understanding of the financial realities of gambling doesnât seem logically inconsistent with deriving enjoyment from itâpeople like winning even when a win has absolutely zero practical import, and people like dopamine hits. Nor does a full understanding seem inconsistent with valuing that enjoyment at $33.50, which is comparable to what they pay for nights of entertainment they report as having roughly equivalent value.
In contrast, Iâm having a hard time characterizing a plausible mechanism of action for the FarmKind scenario that produces some valued internal state (cf. âentertainedâ in the casino example) without either conflicting with other evidence about the standard donorâs preferences (which charity is their favorite) or being inconsistent with the standard donor having a true understanding of the match (that the only real-world difference from the world in which they directly donated to their preferred charity is more money going to the super-effective charity, with an equal reduction of the amount going to their preferred charity).
While I understand the concern that people might only participate in FarmKindâs matching because they misunderstand it, I believe there are those who could find psychological or strategic value in the process even with full understanding. My main point was that, if participants are fully informed and still choose to engage, then any ethical issues are primarily about ensuring transparency. I disagree with the idea that thereâs something inherently wrong with the process itself, as long as thereâs no deception involved.
I suppose you are saying that this is a situation where the facts are such that there should be a presumption of the participant being misled. But, arguendo, if the participant fully understands the process and chooses to engage, is there some residual wrongfulness? Because I would think the remainder of the inquiry is a prudential question of whether it has the effect of raising more funds for the effective charities.
[note that this is not about FarmKind specifically. Itâs a response on whether encouraging the donor to take a deal that is clearly inconsistent with their professed interests is OK as long as the donor fully understands the arrangement]
I think there is a strongâbut not irrebuttableâpresumption that the person fails to fully understand the process. Based on the differences in opinion, even the Forum readership seems to not be having an easy go at that . . . and thatâs a group of intelligent people whose background should make it a lot easier! I also suspect that relatively few people would go through with it once they had a full understanding. The basis for that suspicion is that it doesnât make sense with the userâs professed preferences, and the mechanism by which the user would gain enough psychological or strategic value have not been clearly defined. It might be possible to address those concerns, but likely at the cost of making the fundraising pitch less effective than other approaches.
Iâd only be willing to tolerate a low number of donors who were not fully informed slipping through the cracks. Part of it is that the org is actively creating the risk for donors to be misled. From a Bayesian point of view, a low base rate for full understanding in the pool of people who tentatively plan to donate would mean that the process for testing whether the person has full understanding needs to have an âaân awfully low false-positive rate to meet my standard.[1] So the charity would need a robust process for confirming that the would-be donor was fully informed. Maybe they could ask Jeff and Ben to do a video on why the process was illusory and/âor misleading in their opinion? After all, itâs hard to be fully informed without hearing from both sides.
After all of that, I question whether you are going to have enough donors completing to make the matching fundraiser more effective than one with a more straightforward pitch. Obviously, oneâs results may differ based on the starting assumption of how many potential donors would go ahead with it if fully informed.
Finally, I question whether this approach plays the long game very well. By the end of the understanding-testing process, the hypothetical donor would know that the modus operandi of the matching org is to ask people to engage in illusory matches that the organization knows are inconsistent with advancing said personâs stated preferences. That does not sound like a good strategy for building long-term relationships with donors . . . and long-term relationships are where a lot of the big money comes in.
Sâuppose only 5% of candidateâs have the required full understanding and should pass. If the process for confirming full understanding iâs even 90 percent accurate, I think weâd ââexpect 4.5 correct pass results, 0.5 false fail results, 81 correct fails, and 9 false passes per 100 examinees. In other words, a full 2â3 of the passes wouldnât have the needed full understanding.
I think I may have hidden the question that I was interested in my response.
I understand with your understanding of the motives and interests of the relevant actors, it is unlikely (maybe impossible?) for someone to do the matching process while understanding it fully.
But if that premise is satisfied, for the sake of argument, does that resolve the ethical question for you? Because my main issue was that Ben was suggesting that it would not.
Assuming the fundraiser, object-level charity, and those in privity with them had always acted in an ethically responsible manner with respect to the donor, I think the confirmed-full-understanding premise could resolve the ethical problem.
(Iâm hedging a bit because I find the question to be fairly abstract in the absence of a better-developed scenario. Also, I added the âalways actedâ caveat to clarify that an actor at least generally cannot behave in a manipulative fashion and then remove the resultant ethical taint through the cleansing fire of confirmed-full-understanding. Iâm thinking of scenarios like sexual harassment of employees and students, or brainwashing your kids.)
If the match is completely illusory, such that no well-informed donor would rationally give it any weight, I would start with a fairly strong presumption that the people donating lack an âadequate understanding of the process.â I donât think that presumption is irrebuttable, but I would probably be looking for something like a pop-up that said something like:
We want you to understand that the âmatchâ here is an illusion. Choosing to donate through this site does not result in more money going to animal charities than donating the same amount directly to your favorite charity would. The only meaningful real-world difference between donating through this site and donating the same amount to your favorite charity is that the effective charity receives more money (and your favorite charity receives less) when you donate through this site. Please initial in the box below to document that you understand this.
I would not start with such a presumption (or at least not a strong presumption) where the evidence for counterfactuality was somewhat stronger (cf. GiveWell). The reason is that the hypothetical well-informed donor probably would give some weight to a match even if they concluded it was (say) 25% counterfactual. Therefore, the fact that a donor is proceeding with a matched donation under those circumstances does not call the adequacy of their understanding into question in the same way.
I do see how this could be adversarial or uncooperative. Do X, or else Iâll stop buying medicine for dying kids. What?!
I think doing something like this could make sense and be reasonably cooperative if the marginal costs of donation were sufficiently burdensome to you (and youâd actually donate less if you werenât trying to match). On your own, the cost-effectiveness â the benefits from your donations divided by your costs â might not look so good, but the benefits from matcheesâ donations + the benefits from your donations all divided by your costs could make it seem worth taking on the extra burden to you. Then, itâs actually you and your matchee(s) cooperating to provide a public good: people or animals being better off. Youâre appealing to others to take this on as a joint responsibility.
This is like people being willing to vote to ban cages and crates for farmed animals, but still buying products from caged and crated animals.
Or taxation for welfare for the poor or other public goods.
I do see how this could be adversarial or uncooperative. Do X, or else Iâll stop buying medicine for dying kids. What?!
Right, I feel like itâs easy to not notice this framing, but it feels pretty weird once you do frame it in that way.
I do agree that there are some circumstances under which donation matches make sense, and increasing marginal returns to donations is perhaps one of them (which is not exactly what you said I think? but similar). I just think these circumstances tend to be relatively niche and I donât see how e.g. the FarmKind case is one of them.
One thing Iâd like to add is that when thereâs pushback on donation matching initiatives, the discussion often focuses on âis the communication of what happens clear?â and âare donors misled?â and so on. I think (lack of) honesty and clarity is the biggest problem with most donor matches, but even where those problems are resolved, I think there are still other problems (e.g. it just seems kind of uncooperative /â adversarial to refuse to do something good that youâre willing and able to do unless someone else does what you want). My favoured outcome is not âhonest matchingâ but âno matchingâ, and Iâd like the discussion to at least have that on the table as a possible conclusion.
(To be clear, I think neither Giving Multiplier nor FarmKind are intending to be dishonest, although I think itâs still unclear to me whether they are unintentionally misleading people. I perhaps think that they are applying ordinary standards of truthfulness and care, whereas Iâd like us to have extraordinary standards.)
If providing funds that will contribute to a match has the effect of increasing funds generated to effective charities and is transparent and forthright about the process involved, I donât really see the problem.
It does not seem to me as an uncooperative and adversarial process, but one that gets more people involved in effective giving in a collaborative and fun experience. As the OP suggested, there is a concern that the matcher (as opposed to the fund provider) may be deceived and not realize that the net effect of their participation is less funds to their preferred charity and more to the fund providerâs preferred charity. But if there is adequate understanding of the process, it seems to me that the question boils down to whether or not the availability of Giving Multiplier or Farmkind results in more counterfactual donations going to these charities.
I think this is a pretty interesting aspect of the discussion, and I can see why people would not only agree with this but think it kind of obvious. Here are some reasons why I donât think itâs so obvious:
I think itâs worth noticing when you offer other people something that you wouldnât take yourself. Why is it good for them, but not good for you?
I think you should include your assessment of what value the deal has to someone else in your assessment of whether you have successfully communicated what the deal is. As above, if people are taking an offer that you wouldnât take, and that as far as you see doesnât benefit them, I think you should take that as evidence that you havenât explained the deal (Jason makes this point in another comment)
I think a useful analogy is a casino, say a roulette wheel. The rules of roulette wheels are pretty simple, and completely public. People play them willingly, often believing itâs in their interests to do so. Yet I think roulette wheels are kind of exploitative and even verge on predatory, and I wouldnât feel comfortable running one even to donate the proceeds to charity. Again, opinions can differ here, but for me I see facilitating and encouraging other people to make decisions that you know are bad isnât ethical behaviour, even if they ultimately make the decision freely and willingly. (To be clear, I think offering a donation match is less unethical than running a casino, but I bring up the casino to illustrate how offering people a free and informed choice sometimes still doesnât sit right with me.)
I think your example of the roulette wheel is illustrative⊠People know that they are going to lose money in expectation but it is rational for many of them to play because the expected fun they have from a night at the casino exceeds the expected loss. For example if you churn $500 and the average payback is 92.5%, the expected loss is $33.50 and you might have more fun than with other recreational activities that cost that much. Of course, some people may have gambling problems or may irrationally think that they have an advantage. These are the cases that we should be worried about: people who have issues of understanding or volitional capacity.
Similarly, people can understand whatâs going on with Farmkind or Giving Multiplier and want to participate, understanding that their choice is benefiting the charities that Farmkind or Giving Multiplier prefer.
If they enjoy and understand the process employed, I donât understand why one would think that they were harmed or exploited
On the casino example: I agree that a gambler could decide that the expected financial loss was outweighed by the value they ascribed to the entertainment. However, that doesnât tell us what proportion of people around the roulette wheel fall into that category vs. people with impulse-control disorders or the like vs. people who do not have disorders yet make harmful decisions because they are too influenced by gamblerâs fallacies. I would put the burden on the would-be casino operator to prove that enough[1] of their would-be customers fell into the ârational gamblerâ category before considering Benâs objection to running a roulette wheel have been satisfactorily addressed.
Turning to the analogy, I think there is considerably more evidence of the ârational gamblerâ category existing than the âpeople who truly understand an FarmKind-type match, but enjoy it enough to experience it as a net positive.â A full understanding of the financial realities of gambling doesnât seem logically inconsistent with deriving enjoyment from itâpeople like winning even when a win has absolutely zero practical import, and people like dopamine hits. Nor does a full understanding seem inconsistent with valuing that enjoyment at $33.50, which is comparable to what they pay for nights of entertainment they report as having roughly equivalent value.
In contrast, Iâm having a hard time characterizing a plausible mechanism of action for the FarmKind scenario that produces some valued internal state (cf. âentertainedâ in the casino example) without either conflicting with other evidence about the standard donorâs preferences (which charity is their favorite) or being inconsistent with the standard donor having a true understanding of the match (that the only real-world difference from the world in which they directly donated to their preferred charity is more money going to the super-effective charity, with an equal reduction of the amount going to their preferred charity).
Pondering what âenoughâ means opens up a new can of worms for sure.
While I understand the concern that people might only participate in FarmKindâs matching because they misunderstand it, I believe there are those who could find psychological or strategic value in the process even with full understanding. My main point was that, if participants are fully informed and still choose to engage, then any ethical issues are primarily about ensuring transparency. I disagree with the idea that thereâs something inherently wrong with the process itself, as long as thereâs no deception involved.
I suppose you are saying that this is a situation where the facts are such that there should be a presumption of the participant being misled. But, arguendo, if the participant fully understands the process and chooses to engage, is there some residual wrongfulness? Because I would think the remainder of the inquiry is a prudential question of whether it has the effect of raising more funds for the effective charities.
[note that this is not about FarmKind specifically. Itâs a response on whether encouraging the donor to take a deal that is clearly inconsistent with their professed interests is OK as long as the donor fully understands the arrangement]
I think there is a strongâbut not irrebuttableâpresumption that the person fails to fully understand the process. Based on the differences in opinion, even the Forum readership seems to not be having an easy go at that . . . and thatâs a group of intelligent people whose background should make it a lot easier! I also suspect that relatively few people would go through with it once they had a full understanding. The basis for that suspicion is that it doesnât make sense with the userâs professed preferences, and the mechanism by which the user would gain enough psychological or strategic value have not been clearly defined. It might be possible to address those concerns, but likely at the cost of making the fundraising pitch less effective than other approaches.
Iâd only be willing to tolerate a low number of donors who were not fully informed slipping through the cracks. Part of it is that the org is actively creating the risk for donors to be misled. From a Bayesian point of view, a low base rate for full understanding in the pool of people who tentatively plan to donate would mean that the process for testing whether the person has full understanding needs to have an âaân awfully low false-positive rate to meet my standard.[1] So the charity would need a robust process for confirming that the would-be donor was fully informed. Maybe they could ask Jeff and Ben to do a video on why the process was illusory and/âor misleading in their opinion? After all, itâs hard to be fully informed without hearing from both sides.
After all of that, I question whether you are going to have enough donors completing to make the matching fundraiser more effective than one with a more straightforward pitch. Obviously, oneâs results may differ based on the starting assumption of how many potential donors would go ahead with it if fully informed.
Finally, I question whether this approach plays the long game very well. By the end of the understanding-testing process, the hypothetical donor would know that the modus operandi of the matching org is to ask people to engage in illusory matches that the organization knows are inconsistent with advancing said personâs stated preferences. That does not sound like a good strategy for building long-term relationships with donors . . . and long-term relationships are where a lot of the big money comes in.
Sâuppose only 5% of candidateâs have the required full understanding and should pass. If the process for confirming full understanding iâs even 90 percent accurate, I think weâd ââexpect 4.5 correct pass results, 0.5 false fail results, 81 correct fails, and 9 false passes per 100 examinees. In other words, a full 2â3 of the passes wouldnât have the needed full understanding.
I think I may have hidden the question that I was interested in my response.
I understand with your understanding of the motives and interests of the relevant actors, it is unlikely (maybe impossible?) for someone to do the matching process while understanding it fully.
But if that premise is satisfied, for the sake of argument, does that resolve the ethical question for you? Because my main issue was that Ben was suggesting that it would not.
Assuming the fundraiser, object-level charity, and those in privity with them had always acted in an ethically responsible manner with respect to the donor, I think the confirmed-full-understanding premise could resolve the ethical problem.
(Iâm hedging a bit because I find the question to be fairly abstract in the absence of a better-developed scenario. Also, I added the âalways actedâ caveat to clarify that an actor at least generally cannot behave in a manipulative fashion and then remove the resultant ethical taint through the cleansing fire of confirmed-full-understanding. Iâm thinking of scenarios like sexual harassment of employees and students, or brainwashing your kids.)
If the match is completely illusory, such that no well-informed donor would rationally give it any weight, I would start with a fairly strong presumption that the people donating lack an âadequate understanding of the process.â I donât think that presumption is irrebuttable, but I would probably be looking for something like a pop-up that said something like:
We want you to understand that the âmatchâ here is an illusion. Choosing to donate through this site does not result in more money going to animal charities than donating the same amount directly to your favorite charity would. The only meaningful real-world difference between donating through this site and donating the same amount to your favorite charity is that the effective charity receives more money (and your favorite charity receives less) when you donate through this site. Please initial in the box below to document that you understand this.
I would not start with such a presumption (or at least not a strong presumption) where the evidence for counterfactuality was somewhat stronger (cf. GiveWell). The reason is that the hypothetical well-informed donor probably would give some weight to a match even if they concluded it was (say) 25% counterfactual. Therefore, the fact that a donor is proceeding with a matched donation under those circumstances does not call the adequacy of their understanding into question in the same way.
I do see how this could be adversarial or uncooperative. Do X, or else Iâll stop buying medicine for dying kids. What?!
I think doing something like this could make sense and be reasonably cooperative if the marginal costs of donation were sufficiently burdensome to you (and youâd actually donate less if you werenât trying to match). On your own, the cost-effectiveness â the benefits from your donations divided by your costs â might not look so good, but the benefits from matcheesâ donations + the benefits from your donations all divided by your costs could make it seem worth taking on the extra burden to you. Then, itâs actually you and your matchee(s) cooperating to provide a public good: people or animals being better off. Youâre appealing to others to take this on as a joint responsibility.
This is like people being willing to vote to ban cages and crates for farmed animals, but still buying products from caged and crated animals.
Or taxation for welfare for the poor or other public goods.
Right, I feel like itâs easy to not notice this framing, but it feels pretty weird once you do frame it in that way.
I do agree that there are some circumstances under which donation matches make sense, and increasing marginal returns to donations is perhaps one of them (which is not exactly what you said I think? but similar). I just think these circumstances tend to be relatively niche and I donât see how e.g. the FarmKind case is one of them.