I think these sorts of critiques don’t just apply to EA—it seems to me like just about any intervention would fall into one of them.
AMF-style interventions that focus on specific problems, like malaria nets? As you discuss, these avoid problems 1 and 2 (because they’re doing a specific thing that wasn’t already being done, so they’re not taking away local jobs or displacing local capacity) but are vulnerable to problem 3 (because the specific thing they’re doing may not be what the locals want)
Maybe organizations could avoid problem 3 by setting up a system to get public input on their projects so they can avoid doing projects that locals don’t want? But expand this out, and at that point you’re basically running (part of) a government—after all, aggregating people’s preferences into decisions is essentially what governments do. (After all, “locals” aren’t a homogeneous group with uniform preferences.) And then you definitely run into all the usual problems with preference aggregation, and you certainly are trying to replace (part of) the local government’s role.
Maybe avoid that by working with the local government or local institutions, rather than setting up your own preference aggregation method? Well, if the problem is that the local government and local institutions are bad or corrupt, that’s certainly not a good idea.
Or maybe your intervention could be targeted directly at improving the institutions? In that case you certainly are saying that you know how to run institutions better than the locals, which goes back into the “we know better than you” dynamic. And I thought part of the problem with colonialism was replacing local institutions that were illegible but functional with institutions that looked better to the colonizers, but didn’t actually work well for locals.
Should you hire locals to work for you? Well in that case you’re displacing those locals away from whatever else they would be doing. Don’t hire locals and bring in your own people instead? Then you’re taking away jobs and ignoring local expertise.
It seems like if you wanted an intervention that avoided all of these kinds of potential problems, it would have to have the following properties:
Locals agree that it would be a good idea
Can be done using local resources and expertise
But also:
Not currently being done locally
Not something that local governments should be doing instead
And I’m having a hard time thinking of any intervention that would fit this bill; if the first two things were true, it seems that would imply that locals would do it or ask their governments to do it.
(Even GiveDirectly-style cash transfers could be argued to have the problem: one could argue that local governments should be giving their citizens cash, so that Acemoglu’s critique about “key services of the state being taken over by other entities” would apply here.)
Two other points of this style of critique that I’m confused about are:
If the argument is “it’s bad to give recipients free/cheap X, because that undermines efforts to produce X locally” isn’t that a general argument for protectionism? If instead of a charity providing X, it was a multinational corporation that had found a way to produce X very cheaply—don’t economists usually say in this sort of situation that trade is a good thing, and the multinational should specialize in X while the locals should specialize in something else? Why is it different when it’s a charity involved?
One of the themes of the “abundance discourse” in the US is that a major problem with US infrastructure projects is that they spend too much time listening to various locals and special interest groups, and bending over backwards to avoid even minor side effects, and not enough time just building the things efficiently. Critique (3) here is sort of the opposite of that—it’s saying that projects need to spend more time listening to locals. It’s not obvious here that “listen to locals more” is good—there are real costs to this approach. Of course the cost-benefit is different in different cases (it’s certainly possible that US projects should listen to locals less, and developing country projects should listen to locals more) I’m just saying that it’s not obvious which direction to go.
Maybe organizations could avoid problem 3 by setting up a system to get public input on their projects so they can avoid doing projects that locals don’t want? But expand this out, and at that point you’re basically running (part of) a government—after all, aggregating people’s preferences into decisions is essentially what governments do. (After all, “locals” aren’t a homogeneous group with uniform preferences.) And then you definitely run into all the usual problems with preference aggregation, and you certainly are trying to replace (part of) the local government’s role.
“Preference aggregation” is also what civil society (e.g. associations, free newspapers, labor unions, environmental groups) does. Unless Acemoglu has abandoned social liberalism while I haven’t looked, I am fairly confident he wouldn’t consider all civil society to be “trying to replace (part of) the local government’s role”. So funding civil society is potentially another broad class of interventions that would fit all those desiderata (like @huw’s, it falls under the broader category of “building local capacity”).
Not something that local governments should be doing instead
A broad class of interventions that fit this bill are entrepreneurial interventions. The goal would be to build local capacity temporarily in order to prove out the intervention in the eyes of the government (who will want to see local demand and effectiveness), and then hand it over to them. These can be locally informed or have local adaptations (satisfying (1) and (2)), but wouldn’t necessarily be currently done because they’re new ideas.
it seems that would imply that locals would do it or ask their governments to do it
I think this isn’t true for these kinds of interventions. It’s easy to imagine large gaps between ‘I think this would be a good idea’ and ‘I have the resources, funding, and willingness to do this myself’ or ‘My government will do this’. Governments in particular move slowly and often want proof that an intervention works locally before implementing it themselves.
I think these sorts of critiques don’t just apply to EA—it seems to me like just about any intervention would fall into one of them.
AMF-style interventions that focus on specific problems, like malaria nets? As you discuss, these avoid problems 1 and 2 (because they’re doing a specific thing that wasn’t already being done, so they’re not taking away local jobs or displacing local capacity) but are vulnerable to problem 3 (because the specific thing they’re doing may not be what the locals want)
Maybe organizations could avoid problem 3 by setting up a system to get public input on their projects so they can avoid doing projects that locals don’t want? But expand this out, and at that point you’re basically running (part of) a government—after all, aggregating people’s preferences into decisions is essentially what governments do. (After all, “locals” aren’t a homogeneous group with uniform preferences.) And then you definitely run into all the usual problems with preference aggregation, and you certainly are trying to replace (part of) the local government’s role.
Maybe avoid that by working with the local government or local institutions, rather than setting up your own preference aggregation method? Well, if the problem is that the local government and local institutions are bad or corrupt, that’s certainly not a good idea.
Or maybe your intervention could be targeted directly at improving the institutions? In that case you certainly are saying that you know how to run institutions better than the locals, which goes back into the “we know better than you” dynamic. And I thought part of the problem with colonialism was replacing local institutions that were illegible but functional with institutions that looked better to the colonizers, but didn’t actually work well for locals.
Should you hire locals to work for you? Well in that case you’re displacing those locals away from whatever else they would be doing. Don’t hire locals and bring in your own people instead? Then you’re taking away jobs and ignoring local expertise.
It seems like if you wanted an intervention that avoided all of these kinds of potential problems, it would have to have the following properties:
Locals agree that it would be a good idea
Can be done using local resources and expertise
But also:
Not currently being done locally
Not something that local governments should be doing instead
And I’m having a hard time thinking of any intervention that would fit this bill; if the first two things were true, it seems that would imply that locals would do it or ask their governments to do it.
(Even GiveDirectly-style cash transfers could be argued to have the problem: one could argue that local governments should be giving their citizens cash, so that Acemoglu’s critique about “key services of the state being taken over by other entities” would apply here.)
Two other points of this style of critique that I’m confused about are:
If the argument is “it’s bad to give recipients free/cheap X, because that undermines efforts to produce X locally” isn’t that a general argument for protectionism? If instead of a charity providing X, it was a multinational corporation that had found a way to produce X very cheaply—don’t economists usually say in this sort of situation that trade is a good thing, and the multinational should specialize in X while the locals should specialize in something else? Why is it different when it’s a charity involved?
One of the themes of the “abundance discourse” in the US is that a major problem with US infrastructure projects is that they spend too much time listening to various locals and special interest groups, and bending over backwards to avoid even minor side effects, and not enough time just building the things efficiently. Critique (3) here is sort of the opposite of that—it’s saying that projects need to spend more time listening to locals. It’s not obvious here that “listen to locals more” is good—there are real costs to this approach. Of course the cost-benefit is different in different cases (it’s certainly possible that US projects should listen to locals less, and developing country projects should listen to locals more) I’m just saying that it’s not obvious which direction to go.
“Preference aggregation” is also what civil society (e.g. associations, free newspapers, labor unions, environmental groups) does. Unless Acemoglu has abandoned social liberalism while I haven’t looked, I am fairly confident he wouldn’t consider all civil society to be “trying to replace (part of) the local government’s role”. So funding civil society is potentially another broad class of interventions that would fit all those desiderata (like @huw’s, it falls under the broader category of “building local capacity”).
So for your criteria:
Locals agree that it would be a good idea
Can be done using local resources and expertise
Not currently being done locally
Not something that local governments should be doing instead
A broad class of interventions that fit this bill are entrepreneurial interventions. The goal would be to build local capacity temporarily in order to prove out the intervention in the eyes of the government (who will want to see local demand and effectiveness), and then hand it over to them. These can be locally informed or have local adaptations (satisfying (1) and (2)), but wouldn’t necessarily be currently done because they’re new ideas.
I think this isn’t true for these kinds of interventions. It’s easy to imagine large gaps between ‘I think this would be a good idea’ and ‘I have the resources, funding, and willingness to do this myself’ or ‘My government will do this’. Governments in particular move slowly and often want proof that an intervention works locally before implementing it themselves.
Do you have an example success story of the kind of intervention that you have in mind?