Thank you for writing this article! As a complete newcomer to pandemic preparedness at large, I found this extremely useful and a great example of work that surfaces and questions often unstated assumptions.

Although I don't have enough expertise to provide much meaningful feedback, I did want to bring up some thoughts I had regarding your arguments in Reason 2. Your 44 hospitalizations threshold in the numerical examples strikes me as reasonable, but it does also seem to me that the metagenomic sequencing of COVID-19 was related ― if not a critical precondition ― to China confirming its finding of a novel pathogen (source). I recognize that the early detection interventions you are calling into question here may be more of the form of mass/representative sampling programs, but it seems plausible to me that merely having the means to isolate and sequence a pathogen near the site of the outbreak in question could substantially affect time to confirmation.

My prior is that China is likely quite capable in that regard, but other countries may have fewer capabilities. All this to say that investing in more conventional metagenomic sequencing capacity in "sequencing deserts" could still be very cost effective. But note that this is all conjectural; I don't know anything about the distribution of sequencing capacity, nor even what it takes to identify, isolate and sequence a pathogen.

Thanks again for this brilliant piece!

Thanks for your questions, Moritz! I hadn't seen that post. 

I don't think I could tell you with confidence that such interventions would be promising. There are a lot of steps in the causal chain that I'm unsure about.

1. Looking at the supplementary material of this study, it seems like a lot of the second-hand cages from the EU's ban may have been sold to countries that already almost only use cage-free systems (cross-referencing with OWID data numbers ). So perhaps many of second-hand cages are simply replacing old ones, rather than causing more chickens to end up in them (of course they could be used for new chickens farms ― don't know).
   1. Note that the OWID numbers are more recent, so it's possible some of these countries increased their cage share as a result of those sales.
2. Still, it's possible that more are now being sold to e.g. Nigeria, and that this is placing new chickens in cages. So the question is: how much does the lower cost of second hand cages induce more adoption? I would think some bit, but I find it hard to say.
3. How concentrated is the supply of second-hand cages? Would bans in a few exporting countries/companies restrict the supply a lot? 
4. What systems are they transitioning from, and what are the improvements in benefits?

It also feels like this would maybe be more of delaying tactic, as producers may upgrade once they are wealthier or new cages are cheaper.

Also, from the spreadsheet you shared with me, it seems like a lot of the projected growth in chicken consumption is in Nigeria. The more the consumption growth is concentrated in one or a few countries, the more it may make sense to focus on their policies.

Sorry, I know this isn't really an answer to your questions ― it's hard to think about!

Thanks for writing this! It's a great report. Some quick thoughts:

• Even if the removal of subsidies caused a decrease in the number of local farms, you might worry that it increases the competitiveness of foreign animal products that may have worse welfare standards (of course, removing subsidies would be a notable policy win, and spur removals in other countries).
• I agree that understanding the kinds of systems used and their prevalence seems key. However, I think it's possible that we could define a lower bound of the number of animals in more intensive systems and pencil out whether an intervention seems worth it (there could still be better interventions at other parts of the intensification gradient). Getting better data sounds like it could be expensive is one concern I have.

On another note, I wonder if a better way to slow the expansion of intensive livestock production in new regions might be to restrict access to critical equipment and inputs. Low/middle income countries might not have the industries to produce cages, crates, slaughter equipment, antibiotics, etc, at scale, and so may import. Influencing policymakers in countries that supply these inputs may be more tractable. For instance, it seems the World Bank has been influenced by high-income countries to limit finance for fossil fuel projects.

Voters in high-income countries want action on moral issues, but often struggle to personally swallow the costs. Imposing the policies on other countries can therefore be an easy policy win for politicians. It wouldn't surprise me if bans on exports of battery cages, gestation crates, antibiotics for use in livestock and such were reasonably tractable (note: not sure antibiotic bans would be good for welfare). Of course all this comes with the complexity of what it may do to wellbeing of humans in affected countries.