Thanks for a clear and insightful post—comparing different cause areas is challenging enough, and it’s particularly hard to weigh up different value sets at the same time (i.e. anti-paternalism and measurable utility). And thanks for including my brief comments on GD, although not sure my insights really warrant the concluding line:
“Nonetheless, even if GiveDirectly is less cost-effective than other charities, there may be other reasons to donate to GiveDirectly. One could for example argue, as George Howlett does, that GiveDirectly promises substantial systemic benefits and that its model is a great way to attract more people to the idea of effective charity.”
I’ll flesh that point out a little more detail here. Essentially I think that a huge proportion of GiveDirectly’s impact comes in areas that necessarily go uncaptured by GiveWell’s quantitative approach. GiveWell even give a pretty clear nod of the head to possibilities here (‘our review focuses on its direct impact, rather than the experimentation or policy impact its programs might have’). Systemic influences are hard to estimate, but may be of great significance in this case—here are a couple of mechanisms for how they could work:
Scaling influence and changing development norms: GD’s simple operating narrative provides it dimensions of effectiveness that other top charities do not possess to the same extent:
-The direct cash transfer model is interesting and easy to understand, and has led to GD getting lots of public attention compared to our other causes. After all, trusting the world’s poorest to know what to do with donations is an attractive idea, especially when coupled with hard evidence of cost-effectiveness. So while I see little reason to inherently value anti-paternalism, then the idea of anti-paternalism is a powerful one that may bring effectiveness benefits in terms of increased public support, over and above our other top charities. I’ve certainly experienced this in my corporate outreach work—people feel an affinity with the basic idea of GD more than, say, deworming. Google were convinced by the model, and helped them get off the ground with a $2.4m grant back in 2012. Then again this point is more around future effectiveness (i.e. greater room for widespread support than other top causes) rather than what goes uncaptured now.
-Similarly, there is much talk in the development sphere around seeing cash transfers as a ‘benchmark’ intervention—i.e. if there isn’t a strong case for doing anything else then we should just give people cash (Ban Ki-moon recently made this case). Contributing to what seem like positive shifts in development sector evaluation norms is highly impactful. GD are only part of the influencing force here, but have some strong government links (e.g. Jo Macrae—GD’s new Head of European Partnerships—used to be Head of UK Humanitarian Policy at DfID).
-This point around scalability of influence and broad appeal externalises to a whole range of persuasive approaches. We can use these to expand the movement’s reach to groups who may have more in common with us than is realised, without watering down what GD do at all. For example I see GD as a specifically feminist intervention: UCTs directly empower the heads of households in East Africa, who are typically female. Or for a business-minded audience: UCTs empower microbusinesses in East Africa, and promote the spread of mobile banking technology (n.b. both of these factors may also contribute to reducing the very valid ‘missing markets’ objection you raise). I listed off a load of these in a FB group post if anyone’s interested—this narrative flex is fascinating! [www.fb.com/groups/effective.altruists/permalink/1292702440786110]
[comment 2⁄2 on GD’s uncaptured effectiveness value via systemic influence—separating comments as they raise related but essentially distinct issues)
UBI: GD’s universal basic income experiment is currently the world’s largest. While nobody can really know what effects a nationally implemented a UBI would have, it could be an incredibly effective tool for reducing inequality, unlocking human flourishing, etc. It’s easy to discount the value of the messy and unmeasurable, but if GD’s work hastens the route to nations considering the idea seriously then this could comfortably trump all its other effectiveness benefits (and speaking of Trump—perhaps UBI adoption would reduce the economic fears that nationalist demagogues can exploit, leading to huge positive impacts in everything from aid and trade policy to X-risk concerns). Systemic change is the only way to achieve real global progress, and promoting UBI is a plausibly good bet in such a highly unpredictable sphere. Donating to GD may be the best buy for those who wish to see the idea tested properly.
I can summarise all of this (including my previous comment) into saying that GD may well be a lot more effective than the quantifiables suggest. In other words, I think GD’s potential for systemic influence could well far outweigh the deficit in provable cost-effectiveness that they have vs some of the other top charities. That being said, this is a very uncertain area and I also donate elsewhere.
So I’m in general agreement with your points on discounting anti-paternalism, and am also aware that I may have picked up a slight pro-GD bias as a result of doing a load of CEA corporate outreach work with them recently. But you did mention that some of your conclusions may not hold if we were to relax the assumption that GiveWell’s cost-effectiveness estimates are accurate. While the points I’ve raised around uncaptured value are (quite rightly) not GiveWell’s territory, do they persuade you to relax this assumption somewhat? And would this influence where you might donate?
Should also add that it’s great to see you highlight that our other top charities are also not paternalistic—more noise should be made about this as a lot of people care. More broadly then I’d also love to hear what uncaptured effectiveness impacts our other top charities might be having, as I’m not really comparing like with like in a post such as this. In fact a discussion of uncaptured value probably deserves a full post of its own, led by someone with more evaluation expertise than me!