I’m somewhat sympathetic to something like GiveDirectly’s take. If bednets are something like 10x more valuable than the cash used to purchase them, I find it a bit weird that people don’t usually buy them when given a cash transfer.
I’ve previously written a short comment about mechanisms that could explain this and do think there are important factors that can explain part of the gap (e.g. coordination problems). But I’m still a bit skeptical that the “real” value is 10x different.
I’m sympathetic to that take as well, but after reading psychology, RCT results and living in Uganda for 10 years it doesn’t surprise me at all that the “real” value of a net could be 10x cash, yet people don’t buy them.
Yes your points about lack of information, lack of markets and externalities can explain the gap in part. Also short term thinking is a big problem.
Even just prioritising urgent over important is a massive and understandable issue. For example sending kids to school in Uganda is so highly valued, yet most people can’t send all their kids to school (especially once some reach high school age) so every shilling could be spent there alone. Then there are medical fees, cultural events like funerals and weddings which are highly prioritized, bride prices etc. In this context making the decision to spend 3 to 5 dollars on a mosquito net is no joke.
There’s also the issue now with nets specifically that they are seen as a “free” item, because they normally are, given away every few years. This means people will often hold out and wait to receive them for free rather than buy them even if they do value them quite highly. Only my richest friends here would even think of buying a net almost based on this alone I think, even disregarding other reasons.
There are some great examples in “poor economics” which talk about similar issues as well.
Maybe sending 200 billion dollars a year or whatever of cash would fix all the problems and make these issues obsolete, but between now and 200 billion I’ll probably lean towards nets and other effective interventions one the EA front while encouraging givedirectly to access as much funding they can through all other means possible. I hope they are giving away billions a year within a few years.
Given that Open Philanthropy seems to believe that typical GiveWell recommendations are dominated by more leveraged ones (e.g.using advocacy, induced technological change) at least for risk-neutral donors, I am a bit confused by the anchoring on GiveWell charities.
Even if GD were closer to AMF than GiveWell thinks, this would not put GD close to the best thing one can do to improve human welfare unless one applies a very narrow frame (risk-aversion, highly scalable based on existing charities right now).
Or, put a bit differently:
(1) We live in a world that has lots of mechanisms in it that make it quite unlikely that something like marginal consumption changes is close to the best use of money—stuff like public good provision, negative and positive externalities, other market failures, policies, technologies etc. with global ramifications etc.
(2) Obviously, it is much harder and uncertain to positively shape the trajectory of, say, rapid malaria vaccines, or trade policy, or to positively influence development aid at large, which is why GD is a good fallback for risk-averse donors and a great baseline to exist.
(3) But it seems extremely unlikely that GiveDirectly is anywhere close to the best thing and it also seems far from clear that pushing the narrative more in the direction of “let’s do more cash transfers” is good on net, unless one is quite sure that the benefits in explainability and scaleability outweigh the impact differential on a per dollar basis on finding and funding interventions with higher EV.
E.g. I would not be surprised if we could find 100 GD interventions in the leveraged GHD space and it is not super clear that we could get less than 1/10th of the funding for them as for GD / a cash transfer centric approach.
Love this and I mostly agree with your points. I do think though that GiveWell is the easiest thing to compare to in this case, and that’s probably fair enough. Comparing to very different and harder-to-measure policy and tech work is less easy to understand and feels a bit disparate.
My only tiiiiiiny nicpick would be your point 2 - I don’t think its that hard to positively shape the trajectory of malaria vaccines (although yes trade policy and influencing development aid is hard-ish). The uncertainties are high yes, but especially with malaria vaccines I would hazard a guess that even the lower end of the effect range might compete with GiveDirectly. Can’t be bothered trying to calculate that right now though :D :D :D
I think even for something that seems quite certain on the intervention level (if you think that is true for malaria vaccine) then one needs to account for funding and activity additionality which make this more uncertain and, relatively speaking, lowers the estimate to GD where the large size of the funding gap ensures funding and activity additionality to be near 1 (i.e. no discount).
Great comment, I appreciate this perspective and have definitely updated towards thinking the 10x gap is more explainable than I thought.
I do note that some of the examples you gave still leave me wondering if the families would rather just have the cash. Sure, perhaps it would be spent on high-priority and perhaps social signal-y things like weddings. But if they can’t currently afford to send all their kids to school or other medical treatment, I wonder if they’d sensibly rather have the cash to spend on those things than a bednet.
(Also, my understanding from surveys of cash recipients is that most do spend their money on essentials or investments.)
Oh I’m almost sure they would rather have the cash, I’m not arguing against that. And yes the evidence is clear that they mostly spend it on essentials—I would argue most of the items considered investments you see (iron roofs, motorbikes etc.) are borderline essentials anyway, or at least you could call them an investment in everyday welfare that will quickly pay dividends either in future wellbeing or financially.
Which one is more effective though depends on how much we weight preference. Give Directly (see their comment) weight the preferences of the recipients most heavily “GiveDirectly believes that the weights that should count the most are those of the specific people we’re trying to help” (Givedirectly’s “North star), while GiveWell (and myself) more heavily weight objective measures, like DALYs averted/QALYs gained and longer term financial benefits, which is where mosquito nets will dominate whatever people can spend the equivalent 5 dollars on.
For example that 5 dollars might pay for half a term’s school fees at a crappy village school (which will bring some benefit), wheras sleeping under a net for 2 years might reduce time off school and improve iron levels which helps learning and brain development, while also reducing the chance of dying or serious disability. Not too hard to see the potential 10x benefit. Apologies if this is obvious and I’m sucking eggs here.
I do wonder though how much motivated reasoning comes into give directly’s take on impact. Obviously if we weight pure preference most heavily, cash will dominate everything—perhaps even Strongminds measured by Wellbys ;). I know from experience how hard it is as someone running an NGO not to lean (or even swing) towards measures which will seem to favour your own intervention above others.
In theory people will always prefer cash because they can spend it on whatever they want (unless of course it is difficult to buy what they most want). This isn’t really up for debate.
What is up for debate is if people actually spend money in a way that most effectively improves their welfare. It sounds paternalistic to say, but I suspect they don’t for the reasons Nick and others have given.
While I’m generally sympathetic to GiveDirectly’s position (I really like their work on so many fronts and think that cash outperforms so many interventions), it seems intuitive to me that it often won’t outperform the very best interventions until we have a lot more funding supply (and I applaud their ambition for increasing that funding supply).
I often think of interventions like bednets as analogous to vaccines (something else that is often distributed for free when there’s a widespread disease instead of sold for cash) for a few reasons:
Stopping the spread: Much like vaccines, bednets are not just about protecting an individual but also the wider community. They achieve this by disrupting the lifecycle of malaria parasites and thereby stopping their spread. This is somewhat similar to how vaccines work to achieve herd immunity, protecting even those who are not vaccinated.
Undervalued goods: Both bednets and vaccines are often undervalued by the very people who would benefit from them. This behavioural quirk might explain why people don’t always buy bednets even when given cash. Similarly, vaccines are often more likely to be used when provided for free, underscoring the importance of removing cost barriers to accelerate their uptake.
Economies of scale & availability: When bednets or vaccines are distributed en masse to an entire region, not only do economies of scale make each unit more cost-effective, but availability also increases. If the purchase of these products were to rely solely on individual decisions made by recipients of cash transfers, there might be insufficient demand to justify large-scale supply, potentially leading to lower availability and higher costs per unit. This contrasts with targeted interventions, where bulk purchasing and distribution can ensure both cost-effectiveness and widespread availability.
For these reasons I’m very glad that COVID vaccines were provided for free in my country instead of charged for and people given the equivalent amount of cash.
Cash transfers are a fantastic tool with broad utility but lack the targeted impact that you can get with specific interventions like bednets or vaccines for the reasons above.
I deeply understand the appeal of cash transfers and the autonomy they offer to recipients and think that they are something that can truly scale but feel like there are compelling and reasonably intuitive reasons for why interventions like providing bednets might be more effective at saving and improving lives.
If the purchase of these products were to rely solely on individual decisions made by recipients of cash transfers, there might be insufficient demand to justify large-scale supply, potentially leading to lower availability and higher costs per unit.
One might even say that providing complete autonomy of choice between bednets and cash to recipients is impossible at a constant funding level. If one could do a widespread distribution of bednets for $2 a person in an area, it should often be feasible to distribute $2 (less administrative costs) to each person instead. However, if bednets now cost $3 because the efficiencies of mass distribution have been lost, then “receive a bednet for free” is no longer something the beneficiary can actualize. In other words, without increasing the funding to allow for “choose between $2 cash or a $3 bednet,” replacing bednets with cash doesn’t really give those who prefer bednets the freedom to choose bednets.
I’m somewhat sympathetic to something like GiveDirectly’s take. If bednets are something like 10x more valuable than the cash used to purchase them, I find it a bit weird that people don’t usually buy them when given a cash transfer.
I’ve previously written a short comment about mechanisms that could explain this and do think there are important factors that can explain part of the gap (e.g. coordination problems). But I’m still a bit skeptical that the “real” value is 10x different.
I’m sympathetic to that take as well, but after reading psychology, RCT results and living in Uganda for 10 years it doesn’t surprise me at all that the “real” value of a net could be 10x cash, yet people don’t buy them.
Yes your points about lack of information, lack of markets and externalities can explain the gap in part. Also short term thinking is a big problem.
Even just prioritising urgent over important is a massive and understandable issue. For example sending kids to school in Uganda is so highly valued, yet most people can’t send all their kids to school (especially once some reach high school age) so every shilling could be spent there alone. Then there are medical fees, cultural events like funerals and weddings which are highly prioritized, bride prices etc. In this context making the decision to spend 3 to 5 dollars on a mosquito net is no joke.
There’s also the issue now with nets specifically that they are seen as a “free” item, because they normally are, given away every few years. This means people will often hold out and wait to receive them for free rather than buy them even if they do value them quite highly. Only my richest friends here would even think of buying a net almost based on this alone I think, even disregarding other reasons.
There are some great examples in “poor economics” which talk about similar issues as well.
Maybe sending 200 billion dollars a year or whatever of cash would fix all the problems and make these issues obsolete, but between now and 200 billion I’ll probably lean towards nets and other effective interventions one the EA front while encouraging givedirectly to access as much funding they can through all other means possible. I hope they are giving away billions a year within a few years.
Given that Open Philanthropy seems to believe that typical GiveWell recommendations are dominated by more leveraged ones (e.g.using advocacy, induced technological change) at least for risk-neutral donors, I am a bit confused by the anchoring on GiveWell charities.
Even if GD were closer to AMF than GiveWell thinks, this would not put GD close to the best thing one can do to improve human welfare unless one applies a very narrow frame (risk-aversion, highly scalable based on existing charities right now).
Or, put a bit differently:
(1) We live in a world that has lots of mechanisms in it that make it quite unlikely that something like marginal consumption changes is close to the best use of money—stuff like public good provision, negative and positive externalities, other market failures, policies, technologies etc. with global ramifications etc.
(2) Obviously, it is much harder and uncertain to positively shape the trajectory of, say, rapid malaria vaccines, or trade policy, or to positively influence development aid at large, which is why GD is a good fallback for risk-averse donors and a great baseline to exist.
(3) But it seems extremely unlikely that GiveDirectly is anywhere close to the best thing and it also seems far from clear that pushing the narrative more in the direction of “let’s do more cash transfers” is good on net, unless one is quite sure that the benefits in explainability and scaleability outweigh the impact differential on a per dollar basis on finding and funding interventions with higher EV.
E.g. I would not be surprised if we could find 100 GD interventions in the leveraged GHD space and it is not super clear that we could get less than 1/10th of the funding for them as for GD / a cash transfer centric approach.
Love this and I mostly agree with your points. I do think though that GiveWell is the easiest thing to compare to in this case, and that’s probably fair enough. Comparing to very different and harder-to-measure policy and tech work is less easy to understand and feels a bit disparate.
My only tiiiiiiny nicpick would be your point 2 - I don’t think its that hard to positively shape the trajectory of malaria vaccines (although yes trade policy and influencing development aid is hard-ish). The uncertainties are high yes, but especially with malaria vaccines I would hazard a guess that even the lower end of the effect range might compete with GiveDirectly. Can’t be bothered trying to calculate that right now though :D :D :D
I think even for something that seems quite certain on the intervention level (if you think that is true for malaria vaccine) then one needs to account for funding and activity additionality which make this more uncertain and, relatively speaking, lowers the estimate to GD where the large size of the funding gap ensures funding and activity additionality to be near 1 (i.e. no discount).
Great comment, I appreciate this perspective and have definitely updated towards thinking the 10x gap is more explainable than I thought.
I do note that some of the examples you gave still leave me wondering if the families would rather just have the cash. Sure, perhaps it would be spent on high-priority and perhaps social signal-y things like weddings. But if they can’t currently afford to send all their kids to school or other medical treatment, I wonder if they’d sensibly rather have the cash to spend on those things than a bednet.
(Also, my understanding from surveys of cash recipients is that most do spend their money on essentials or investments.)
Oh I’m almost sure they would rather have the cash, I’m not arguing against that. And yes the evidence is clear that they mostly spend it on essentials—I would argue most of the items considered investments you see (iron roofs, motorbikes etc.) are borderline essentials anyway, or at least you could call them an investment in everyday welfare that will quickly pay dividends either in future wellbeing or financially.
Which one is more effective though depends on how much we weight preference. Give Directly (see their comment) weight the preferences of the recipients most heavily “GiveDirectly believes that the weights that should count the most are those of the specific people we’re trying to help” (Givedirectly’s “North star), while GiveWell (and myself) more heavily weight objective measures, like DALYs averted/QALYs gained and longer term financial benefits, which is where mosquito nets will dominate whatever people can spend the equivalent 5 dollars on.
For example that 5 dollars might pay for half a term’s school fees at a crappy village school (which will bring some benefit), wheras sleeping under a net for 2 years might reduce time off school and improve iron levels which helps learning and brain development, while also reducing the chance of dying or serious disability. Not too hard to see the potential 10x benefit. Apologies if this is obvious and I’m sucking eggs here.
I do wonder though how much motivated reasoning comes into give directly’s take on impact. Obviously if we weight pure preference most heavily, cash will dominate everything—perhaps even Strongminds measured by Wellbys ;). I know from experience how hard it is as someone running an NGO not to lean (or even swing) towards measures which will seem to favour your own intervention above others.
In theory people will always prefer cash because they can spend it on whatever they want (unless of course it is difficult to buy what they most want). This isn’t really up for debate.
What is up for debate is if people actually spend money in a way that most effectively improves their welfare. It sounds paternalistic to say, but I suspect they don’t for the reasons Nick and others have given.
While I’m generally sympathetic to GiveDirectly’s position (I really like their work on so many fronts and think that cash outperforms so many interventions), it seems intuitive to me that it often won’t outperform the very best interventions until we have a lot more funding supply (and I applaud their ambition for increasing that funding supply).
I often think of interventions like bednets as analogous to vaccines (something else that is often distributed for free when there’s a widespread disease instead of sold for cash) for a few reasons:
Stopping the spread: Much like vaccines, bednets are not just about protecting an individual but also the wider community. They achieve this by disrupting the lifecycle of malaria parasites and thereby stopping their spread. This is somewhat similar to how vaccines work to achieve herd immunity, protecting even those who are not vaccinated.
Undervalued goods: Both bednets and vaccines are often undervalued by the very people who would benefit from them. This behavioural quirk might explain why people don’t always buy bednets even when given cash. Similarly, vaccines are often more likely to be used when provided for free, underscoring the importance of removing cost barriers to accelerate their uptake.
Economies of scale & availability: When bednets or vaccines are distributed en masse to an entire region, not only do economies of scale make each unit more cost-effective, but availability also increases. If the purchase of these products were to rely solely on individual decisions made by recipients of cash transfers, there might be insufficient demand to justify large-scale supply, potentially leading to lower availability and higher costs per unit. This contrasts with targeted interventions, where bulk purchasing and distribution can ensure both cost-effectiveness and widespread availability.
For these reasons I’m very glad that COVID vaccines were provided for free in my country instead of charged for and people given the equivalent amount of cash.
Cash transfers are a fantastic tool with broad utility but lack the targeted impact that you can get with specific interventions like bednets or vaccines for the reasons above.
I deeply understand the appeal of cash transfers and the autonomy they offer to recipients and think that they are something that can truly scale but feel like there are compelling and reasonably intuitive reasons for why interventions like providing bednets might be more effective at saving and improving lives.
One might even say that providing complete autonomy of choice between bednets and cash to recipients is impossible at a constant funding level. If one could do a widespread distribution of bednets for $2 a person in an area, it should often be feasible to distribute $2 (less administrative costs) to each person instead. However, if bednets now cost $3 because the efficiencies of mass distribution have been lost, then “receive a bednet for free” is no longer something the beneficiary can actualize. In other words, without increasing the funding to allow for “choose between $2 cash or a $3 bednet,” replacing bednets with cash doesn’t really give those who prefer bednets the freedom to choose bednets.