Is region-level cause prioritization research valuable to spot promising long-term priority causes worldwide?
post by AmAristizabal (AmAristizábal)
Hi, thanks to a post that I recently wrote about geographic diversity in EA [EA · GW]I have discussed this question with people from other local groups that have more experience in local prioritization and it has led to very interesting conversations (I hope we will have something similar for South America soon). So here are my thoughts:
Choosing the most promising causes is a promising and neglected cause by itself, since there can be large differences in value between causes and the area has received little attention outside EA. So far, there have been more efforts to prioritize specific interventions with more immediate goals by comparing their effectiveness but this narrow view can lead us to ignore other areas that could bring more welfare in the long term even if those are harder to assess and difficult to measure. The value of research in this field (as stated on the cause prioritization overview by 80,000 hours) will come from learning how to build the infrastructure for much better prioritization in the future. Exploring ways of prioritizing causes can be valuable even if the methodologies that are put to the test show to be inadequate, since spotting the difficulties that arise during the process and learning how certain tools are inaccurate would facilitate future efforts and shed light on the weaker spots
Existing organizations that develop cause prioritization research usually evaluate priorities from a global perspective; however, I wonder if cause neutrality is necessarily against local prioritization. Perhaps there is a scenario in which we can maintain cause neutrality by having a portfolio of regional causes instead of (or parallel to) a list of global ones. Also, it is not always clear that it would be more cost-effective to coordinate different actors across the globe concerned with the same causes than to coordinate agents with different preference orderings.
Some groups in lower and middle income countries have tried to come up with their own list of priority causes as a way to give advice to local donors about their comparative advantages and some of the identified cause areas are surprisingly different from other global priority causes in EA. The causes are not necessarily immediate nor specific, and some of them –even if narrow in their geographical scope– can have a relatively long-term scope. I wonder if these efforts could inform global prioritization as a whole if done more broadly across the world. Not only their perceptions on what should be prioritized have informational value, but also their assessment of their comparative advantages deserve more investigation.
Finding priorities depends on reliable comparisons with common metrics, but the search for common metrics requires more abstractions and assumptions about long run effects than narrower prioritization. I wonder if we could be narrower in our geographical scope as a way to counter our broad long-term approach; also it could allow us to develop frameworks to aggregate local priority cause research to spot promising areas. Perhaps this could provide a way to break the broader question into pieces and even if we discover that it leads us in a wrong direction, it could help build better methodologies for future explorations.
Comments sorted by top scores.
comment by Ramiro ·
2020-07-25T13:03:47.181Z · EA(p) · GW(p)
Discussions over local vs. global remind me the contrast between the performances of two Give Directly programs, 100+ (cash tranfers for American families), which received US$ 114.3 mi, and Covid-19 Africa, which received US$ 53.7. I can see reasons for GD supporting 100+, and I'm not surprised that US$1 is more likely to be donated to poor Americans than to sub-saharian Africa, but this made me (and other people, of course, but I speak for me) wonder if we can draw a line between "we're using parochialism to promote EA-like goals" and "we're compromising with parochialism, diverting scarce resources and giving up effectiveness"? I don't think of this as a main issue, but as a puzzle; it would be interesting to have some research on public criteria or clues about this difference.Replies from: Prabhat Soni, AmAristizábal
↑ comment by Prabhat Soni ·
2020-07-26T19:22:39.340Z · EA(p) · GW(p)
I think this is a very relevant point. I think (correct me if I'm wrong) the effectiveness of the best intervention in the world >>> the effectiveness of the best intervention in a random country X. So, it would be more beneficial to have 100 donors for effective global issues compared to 500 donors for effective national issues.
A caveat, however is value promotion. This is difficult to measure or quantify. There is a chance of large spillover effects due more people having an "effective giving" mindset. These people may further spread the idea of effective giving, or may become globally-aligned in the future. Off the top of my head, I think the spillover effects would be rather modest, but we'd probably need more "hard evidence" for this argument.
↑ comment by AmAristizabal (AmAristizábal) ·
2020-07-26T19:26:48.057Z · EA(p) · GW(p)
Definitely. It is a puzzle that I constantly have in mind. I would say that the line could be drawn only when it is used kind of as a “last resource”? Haha so it makes sense to “use parochialism to promote EA-like goals” (and in your example I suppose that not having the 100+ option wouldn´t have meant more funds for Covid-19 Africa) but it makes sense only if there is no possible way to fight parochialism (or if it is excessively costly, which I think is in many contexts and with certain individuals). But as you say, it would be interesting to find where that threshold is (when is it unnecessarily hard to fight parochialism and should we aim for more cost-effectiveness within that restricted scope?). Thanks for the comment!
comment by brb243 ·
2020-07-25T18:13:42.061Z · EA(p) · GW(p)
Research of the most cost-effective causes, interpreted as means to create additional impact can inform long-term priorities – in regions of any levels of development. E. g. in Lokoja in Northern Nigeria, that means may be very different from that in Bangkok or Washington, D. C. Maybe in Lokoja that is informing mothers on the available prenatal and early childhood healthcare incentives (that in the long term gives rise to institutions perpetuating increased wellbeing), in Bangkok supporting regional norms on migrant work, and in Washington, D. C. lobbying for trade policy favorable to LMICs.
Different locally-identified measures can be globally compared in their cost-effectiveness, complementarities potentially concluded, and individual EAs may decide, based on their expertise and the extent of the fulfillment of care of more inner moral circles, whether they wish to focus on a local or more distant measure, or even relocate.
For this global cost-effectiveness comparison and insights into complementarities, knowledge of the entire field of possible impact, as well as the global structure within which the intervention extends and cascades impact, is needed.
Research of the most cost-effective local causes, interpreted as means to help locals, may also inform long-term priorities – also in regions of any levels of development. First, comparison can show where a local should allocate their focus to help most effectively (e. g. a person in Washington, D. C. can conclude that supporting migrant laborers in Southeast Asia is more cost-effective than supporting local homeless persons). Second, complementarities can be also drawn - e. g. a DC-based person may be able to benefit from focusing on a positive measure (e. g. migrant labor laws in Southeast Asia) as opposed to negative-emotions based advertisement - and person in Bangkok can benefit from increased ability to institutionalize positive change. Third, the identification of cost-effective means to help locals enables persons to fulfill their need to care for more inner moral circles [EA · GW] more cost-effectively, so that further funds are left for more outer moral circles.Replies from: AmAristizábal
↑ comment by AmAristizabal (AmAristizábal) ·
2020-07-26T06:05:12.576Z · EA(p) · GW(p)
Thanks for your comment, this is insightful. I like the distinction between as a means to create additional impact and as means to help locals. Also thanks for pointing out other ways in which this latter option informs long-term priorities, there are many I did not consider before such as the further funds that are left for outer moral circles when people care for inner moral circles more cost-effectively, although I wonder if that is always the case or if the time/effort invested finding cost-effective local causes to care for inner moral circles could be better used otherwise, like finding ways to expand moral circles for example haha but so far I share your views and I think that it is valuable to spot these areas acknowledging the limits. Thanks again!
comment by Ramiro ·
2020-07-25T12:59:29.820Z · EA(p) · GW(p)
Thanks for the post! Would you have any examples of causes that could be a local priority, but not a global one?Replies from: BrianTan
↑ comment by BrianTan ·
2020-07-26T12:37:01.450Z · EA(p) · GW(p)
A good example is the criminal justice reform work that Open Phil funds, which is a local priority in the U.S. but not a global one. It's especially pressing in the U.S. given they have the highest incarceration rate in the world.
comment by Prabhat Soni ·
2020-07-26T20:52:32.162Z · EA(p) · GW(p)
I think your question is: Is there some problem/intervention that is high-impact that EA has missed out because it is specific to my country, and so nobody has thought of it?
Let's go through which countries are good for specific causes:
- Artificial General Intelligence: USA, China, UK
- Engineered Pandemics: USA, China
- Earning-to-give: rich countries like USA, Qatar, Singapore, Norway, UAE, Luxembourg, Saudi Arabia, Switzerland
- Nuclear Security: Russia, USA, North Korea
- Climate Change: Countries developing rapidly like Brazil, India and countries that emit a lot of greenhouse gases as of now like USA, UK, etc
- Improving Institutional Decision Making: Corrupt countries like Colombia, Brazil, India Mexico, Ghana, Bolivia and influential countries like USA, UK
- Malaria Interventions: A lot of the countries in Sub-Saharan Africa
- Influencing long-term future: Potential superpowers like Russia, China, India, Brazil
- Alternative meats: Brazil, China, USA, Israel, India
- Food/Water Fortification: India, West African countries
The countries that are good for specific problems/interventions are good because they exhibit certain "structural" properties. For example, countries good for earning to give are rich; countries good for factory farming have high consumption of meat; countries good for institutional decision making are corrupt or influential; countries good for influencing long-term future are potential superpowers; and so on.
These "structural" properties are present in multiple (on average around 5) countries, and thus there are around 5 countries that are high-impact for a specific cause area/intervention. Also, these countries are generall geographically and culturally dispersed - often belonging to different continents.
Coming back to the original question: Is there some problem/intervention that is high-impact that EA has missed out because it is specific to my country, and so nobody has thought of it?
If what I have argued above is correct, the premise that "a problem/intervention is specific to my country" is generally false. Going by the trend that the top ~10 problems/interventions today are not region-specific, I see no reason why a very promising problem/intervention would be found that is region-specific. And, so I argue that region-level cause prioritization research is not particularly valuable.
EDIT: I'm proposing that a majority of the promising problems are not restricted to a particular region. Ofcourse, there are some exceptions to this like war, US immigration, (maybe) health development in Sub-saharan Africa, etc.Replies from: Linch, brb243
↑ comment by Linch ·
2020-07-30T03:10:37.124Z · EA(p) · GW(p)
Some other examples of stuff that seem to benefit a lot from local knowledge:
1. Source control of epidemics/pandemics. A DRC native is presumably much better equipped to understand and deal with the 2020 Ebola outbreak than I am.
2. Wars, especially local wars, H/T brb243 [EA · GW]
3. Improving Institutional decision-making in the "laboratories of democracy" sense. I think part of the value is the direct impact of reducing corruption, etc, but a lot of it is having sufficient experimentation with local politics and then being able to copy over the lessons to other larger, more influential, governments. For example, you can imagine that if Europe's GDPR or Brazil's LGPD are good ideas, other countries will copy the better ones over. Similar stories may be true for micro-experiments in congestion pricing or pandemic preparedness. That said, the Western world's bizarre unwillingness to listen to East Asia about the current pandemic undercuts my point a lot.
4. Local animal activism. I think (medium-low confidence) there are a lot of optics and logistical issues with outsiders being overly "pushy" about animal activism, and it's usually better for such things to arise mostly organically from within. Replies from: AmAristizábal
↑ comment by brb243 ·
2020-07-29T14:40:37.327Z · EA(p) · GW(p)
What about the Cameroonian Civil War that (or at least of which effects) can be mitigated by a combination of EA and local knowledge? This can be a potentially high-impact problem/intervention that has not been covered by other EA research, perhaps due to its localized nature.Replies from: Prabhat Soni