Excellent and underrated post. I actually told Greg a few years ago that this has become part of my cognitive toolkit and that I use this often (I think there are similarities to the Tinbergen Rule—a basic principle of effective policy, which states that to achieve n independent policy targets you need at at least n independent policy instruments).
This tool actually caused me to deprioritize crowdfunding with Let’s Fund, which I realized was doing a multiobjective optimization problem (moving money to effective causes and doing research), and that I needed to focus on one thing.
Another instance in which I used this is in my climate policy paper, where I mentioned suspicious convergence:
“Advanced economies sometimes give aid to emerging economies for environmentally harmful projects: to increase tourism,[278] to build gas power plants,[279] and sometimes even to build coal power plants.[280] Does this reflect a lack of policy coherence? Why not fund projects that make sense from both the perspective of the climate and poverty reduction? For instance, one natural experiment in Brazil showed that paying “extremely poor households for forest conservation” reduced deforestation by 3-5%.[281] A recent randomized controlled trial[282] found that conditional cash transfers to forest-owning Ugandan farmers to conserve forest owned by them prevented emissions at a rate of $0.46 per ton of CO₂.[283]
Given that the lower bound for the social cost of carbon has been estimated to be $125 per ton,[284],[285] should there be a scaling up of such interventions? Would this policy-coherent approach to preventing both poverty and climate change be the most effective? We argue that—perhaps counterintuitively—it might not be.
At a first approximation, a policy-coherent approach appears preferable, and giving aid for gas plants seems counterproductive. However, gas will make up a non-trivial fraction of energy for the foreseeable future, and, in terms of emissions and air pollution it produces, gas is much better than coal. Energy access is vital for industrial development, which reduces poverty; despite the fact that it violates principles of policy coherence, it might be optimal to give aid for gas power. To get a bit more technical: Multi-objective optimization is generally harder than single-objective optimization.[286] It might therefore be more effective to optimize for poverty reduction or economic growth in aid project A, be that through fostering tourism or cheaper electricity access through gas. Then , in ‘aid’ project B (which then is not really an aid project, but a climate change project), one should optimize for the most effective climate change prevention . There is an allure to policy coherence and optimizing for several objectives at once, but it would be a suspicious convergence if the best poverty reduction methods happened to be the most effective ways to combat climate change as well.
Aid should reduce poverty and/or stimulate growth at the same time that other funding is used to combat climate change in the most effective way. One of these ways is performance-based pay for the conservation of rainforests.[287] For example, Norway pledged up to $1 billion in performance-based pay for the conservation of Brazilian rainforests. Preventing deforestation in Brazil is very cost-effective, at $13 per ton of CO₂ equivalent averted on average.[288] Performance targets can be independently verified through satellite imaging. Such efforts should be disconnected from aid budgets, while the international community could pay for the preservation of rainforests globally.”
Excellent and underrated post. I actually told Greg a few years ago that this has become part of my cognitive toolkit and that I use this often (I think there are similarities to the Tinbergen Rule—a basic principle of effective policy, which states that to achieve n independent policy targets you need at at least n independent policy instruments).
This tool actually caused me to deprioritize crowdfunding with Let’s Fund, which I realized was doing a multiobjective optimization problem (moving money to effective causes and doing research), and that I needed to focus on one thing.
Another instance in which I used this is in my climate policy paper, where I mentioned suspicious convergence:
“Advanced economies sometimes give aid to emerging economies for environmentally harmful projects: to increase tourism,[278] to build gas power plants,[279] and sometimes even to build coal power plants.[280] Does this reflect a lack of policy coherence? Why not fund projects that make sense from both the perspective of the climate and poverty reduction? For instance, one natural experiment in Brazil showed that paying “extremely poor households for forest conservation” reduced deforestation by 3-5%.[281] A recent randomized controlled trial[282] found that conditional cash transfers to forest-owning Ugandan farmers to conserve forest owned by them prevented emissions at a rate of $0.46 per ton of CO₂.[283]
Given that the lower bound for the social cost of carbon has been estimated to be $125 per ton,[284],[285] should there be a scaling up of such interventions? Would this policy-coherent approach to preventing both poverty and climate change be the most effective? We argue that—perhaps counterintuitively—it might not be.
At a first approximation, a policy-coherent approach appears preferable, and giving aid for gas plants seems counterproductive. However, gas will make up a non-trivial fraction of energy for the foreseeable future, and, in terms of emissions and air pollution it produces, gas is much better than coal. Energy access is vital for industrial development, which reduces poverty; despite the fact that it violates principles of policy coherence, it might be optimal to give aid for gas power. To get a bit more technical: Multi-objective optimization is generally harder than single-objective optimization.[286] It might therefore be more effective to optimize for poverty reduction or economic growth in aid project A, be that through fostering tourism or cheaper electricity access through gas. Then , in ‘aid’ project B (which then is not really an aid project, but a climate change project), one should optimize for the most effective climate change prevention . There is an allure to policy coherence and optimizing for several objectives at once, but it would be a suspicious convergence if the best poverty reduction methods happened to be the most effective ways to combat climate change as well.
Aid should reduce poverty and/or stimulate growth at the same time that other funding is used to combat climate change in the most effective way. One of these ways is performance-based pay for the conservation of rainforests.[287] For example, Norway pledged up to $1 billion in performance-based pay for the conservation of Brazilian rainforests. Preventing deforestation in Brazil is very cost-effective, at $13 per ton of CO₂ equivalent averted on average.[288] Performance targets can be independently verified through satellite imaging. Such efforts should be disconnected from aid budgets, while the international community could pay for the preservation of rainforests globally.”